Symbiotic and aposymbiotic Aiptasia belonging to the CC7 clonal line (Sunagawa et al., 2009) were reared in 1-L polystyrene tanks filled with 700 mL of artificial seawater (ASW) (Red Sea Coral Pro Salt) at 27°C while being fed Artemia nauplii and supplied with fresh ASW twice per week. Symbiotic Aiptasia were raised in an incubator on a 12 h light:12 h dark cycle subjected to 25 μmol photons m-² s-¹ of white LED light, while aposymbiotic Aiptasia, previously derived from 3-(3,4-dichlorophenyl)-1,1-dimethylurea (DCMU) treatment according to established methods (Xiang et al., 2013), remained in the dark. Ten days prior to the beginning of each experiment, anemones allocated for the experiment were deprived of food to ensure Artemia were not present in the anemones’ gastric cavities.

Two different temperature studies were conducted, each with an ambient treatment at 27°C, and a heat-stress treatment at 34°C, which prompts bleaching (Bieri et al., 2016; Esherick et al. in revision). A preliminary 10-day endpoint study with triplicate replication was performed to assess variance. Following this experiment, a subsequent 8-day time-series experiment was conducted without replication, but with much greater numbers of polyps per time point and associated observation of bleaching. This experiment, which added sequencing of eukaryotic 18S rRNA, was designed to give distinct but complementary results that could be intercompared with the endpoint experiment.

Endpoint experiment. Thirty symbiotic and 30 aposymbiotic polyps were split into twelve separate 700-mL tanks, each containing 5 individuals. Half of the tanks were placed in a 34°C incubator while the other half were kept at 27°C for the entire duration of the 10-day trial. On the final day, triplicate samples, each consisting of a mixture of 5 individual polyps, were collected for DNA extraction ( Figure 1A ). Triplicate samples of seawater were also filtered for DNA extraction for seawater bacterial communities, although only one was sequenced due to cost limitations.

Time-series experiment. Sixty aposymbiotic and 60 symbiotic anemones were incubated at an ambient temperature of 27°C, with four tanks from each symbiotic state harbouring 15 individuals each ( Figure 1B ). On day 0, half of the tanks were moved to a heat-stress temperature of 34°C for the remainder of the 8-day trial. Throughout the 8-day time series, algal fluorescence was used as a qualitative proxy for algal abundance as anemones bleached during thermal stress. Five images per anemone were taken at each time point using a Leica MC190 HD camera equipped to Leica M165FC fluorescence stereo microscope with a GFP longpass filter. All images were standardized using 125 ms exposure, 3.4X magnification, and 10X gain.

DNA samples were collected from Aiptasia and their surrounding seawater throughout the 8-day period. In ambient-temperature treatments, samples were collected on day 0, 4 and 6 ( Figure 1C ). In the heat-stress treatment, samples were collected on days 1, 3, 5, and 7 to observe community responses of both prokaryotes and eukaryotes before and after peak bleaching ( Figure 1C ).

The initial 10-day endpoint experiment, which assessed Aiptasia bacterial microbiomes, yielded 2,508 bacterial OTUs. However, even though each replicate was a mixture of 5 anemones, they showed huge variance, with a minority of OTUs present in all replicates. Alpha diversity, as measured by the Shannon-Weaver index, revealed a slightly significant change due to symbiotic state only in the 34°C treatment ( Figure 3 , p=0.02). A PERMANOVA test showed no significant effect of heat stress on microbiome composition, but a modestly significant effect of symbiotic state ( Table 1 , p=0.01).

The subsequent 8-d time-series experiment took a complementary approach. We used twice as many polyps and merged 15 anemones per time point to observe changes during and after bleaching. We also added sequencing of eukaryotes and archaea. This experiment yielded considerably more bacterial OTUs (11,787), likely due to increased quantity of Aiptasia DNA extracted. Although it lacked replication, this experiment showed consistent patterns across time points and in response to bleaching. Rhodobacteraceae and Alteromonadaceae dominated Aiptasia bacterial microbiomes in both experiments, but the first experiment showed a higher fraction of low-abundance families (40-60%) due to variance, as well as possible contamination, so we will focus on the time-series results here, as they contain more detailed information.

Across algal-symbiotic states and temperature treatments in both Aiptasia and seawater, 24-40% of all OTUs belong to bacterial families Alteromonadaceae and Rhodobacteracaeae ( Figure 4 ). At the start of the experiment prior to temperature shift (day 0), we observed that symbiotic state determined bacterial microbiome composition in Aiptasia ( Figure 4A ) but not the surrounding seawater ( Figure 4B ). Symbiotic Aiptasia were dominated by members of Pelobacteraceae (30%) and Rhodobacteraceae (17%), while aposymbiotic Aiptasia were dominated by members of Alteromonadaceae (22%), with smaller fractions of Rhodobacteraceae (10%) and Kordiimonadaceae (9%), a group that was <1% of the microbiome in symbiotic anemones ( Figure 4A ). In contrast, bacterial communities of the seawater surrounding symbiotic and aposymbiotic animals were less distinct ( Figure 4B ), and were dominated by members of Alteromonadaceae and Rhodobacteraceae, but with significant contributions from members of Sphingomonadaceae and Chitinophagaceae, two sugar-consuming families.

We observed a decrease in chlorophyll fluorescence in heat-stressed Aiptasia, occurring most dramatically between days 3 and 5, most notably in the tentacles ( Figure 5 ). In contrast, fluorescence remained constant in anemones that remained at ambient temperature ( Figure 5 ). 18S sequencing of Aiptasia and seawater revealed consistent shifts in the eukaryotic community over this event. Aiptasia samples were dominated, as expected, by anemone and Symbiodiniaceae DNA, with only traces of other eukaryotes. In seawater, we found little Symbiodiniaceae DNA at day 0 at 27°C, as expected. Diatoms (Nitschzia sp.) were detected on day 0, but they rapidly declined by day 3 ( Figure 6A ), likely due to grazing by a gastrotrich (Chaetonotus), which increased on day 4 ( Figure 6B ). Exclusive to the 34°C treatments, we observed an increase of Symbiodiniaceae (normalized to Aiptasia OTU counts) in seawater from days 3-5, corresponding to Aiptasia bleaching ( Figure 6C ). This declined by days 6-7, following a second bloom of the gastrotrich. Curiously, though, counts of Symbiodiniaceae OTUs in Aiptasia (normalized to Aiptasia counts) did not decrease in the 34°C treatments ( Figure 6D ).

In 27°C samples over 8 days, we observed variability, but no consistent shifts in bacterial microbiomes for Aiptasia, which retained the symbiotic signatures of day 0 ( Figure S2A ). In contrast, seawater samples showed clear shifts. In seawater surrounding symbiotic anemones, the fraction of Rhodobacteraceae and Saprospiraceae members increased, while those of Alteromonadaceae, Sphingomonadaceae and Chitinophagaceae decreased ( Figure S2B ). Even more dramatic was a greater representation of Alteromonadaceae in seawater surrounding aposymbiotic anemones on days 4 and 6, which dominated the community ( Figure S2B ). Reflecting this, alpha diversity indices remained fairly consistent throughout the experiment, aside from the seawater associated with aposymbiotic Aiptasia ( Table 2 ), which decreased markedly.

Under heat stress (34°C), we observed microbiome changes in both symbiotic states of Aiptasia. Just one day following heat stress, symbiotic anemones lost almost all their Pelobacteraceae. This loss extended into bleaching, but recovered after day 5 ( Figure 7A ), along with more Saprospiraceae members present. Seawater from symbiotic anemones also showed a slightly greater abundance of Saprospiraceae following heat stress ( Figure 7B ), particularly around days 3-5 when bleaching was observed, with a corresponding decline of Alteromonadaceae members.

The aposymbiotic anemones, which started with a different microbiome as explained above, also showed strong changes, with an initial larger then smaller fraction of Alteromonadaceae ( Figure 7A ). Most striking was the larger proportion of community members that were barely present at the start of the experiment (‘others’), increasing to over 60% of the total microbiome by day 7 ( Figure 7A ). As with symbiotic anemones, the microbial community in the surrounding seawater reflected these changes, with the proportion of Alteromonadaceae diminishing over time ( Figure 7B ) and a greater representation of ‘others’ to about 40% of the total by days 5-7. Although day 4 of the aposymbiotic anemones at 27°C also showed a large burden of ‘others’ ( Figure S2A ), there was no consistent trend, and ‘others’ in most samples were <15% of the communities.

We also detected several archaeal families, but only in seawater, with one possible exception ( Figure 8 , day 4 of the aposymbiotic Aiptasia at 27°C, but likely a contaminant). The most abundant OTUs resembled Aciduliprofundum, a thermophilic euryarchaeote associated with deep-sea hydrothermal vents, but only at about 85% similarity, with another group mapping to Methanobacteriaceae. Although present at both temperatures and in both symbiotic states, we observed a dramatic increase of both groups in seawater of aposymbiotic anemones following heat stress, which remained high for the duration of the experiment ( Figure 8 ). To our knowledge, this is the first observation of archaea associated with Aiptasia, albeit only in surrounding seawater.

Given that Pelobacteraceae, a family containing potential nitrogen fixers, was present only in symbiotic anemones and responded dramatically to heat stress, we further examined the nitrogen-fixation potential of diverse bacteria associated with the alga-symbiotic state of Aiptasia. To determine the presence of the nifH genes, we tested four degenerate primer sets universal for known nifH genes ( Table S1 ), selecting Ueda19F/R6 as the optimal primer set with an amplicon size of ~454 bp ( Figure S1 , lanes 14-16). These primers amplify with known Pelobacter nifH genes ( Figure S3 ). Sequencing of nifH from root-nodule samples confirmed the diazotroph, Rhizobium leguminosarum, to be the bacterial symbiont of fava-bean plants, as expected. Furthermore, sequencing confirmed the amplification of nifH sequences in symbiotic and aposymbiotic Aiptasia, and multiple diazotroph species were apparent in these samples (Phillips et al., unpublished). nifH expression, assessed via qPCR using cDNA from Aiptasia, showed no difference between symbiotic and aposymbiotic anemones, as nifH was barely expressed in either ( Figure 9A ). To detect whether nifH genes were even present, we attempted qPCR of gDNA from Aiptasia, but the primer efficiencies were out of range when using gDNA from all samples. However, gel analysis of nifH-PCR products using a dilution series from anemone gDNA suggested detectable nifH in Aiptasia microbiomes, with more possibly in aposymbiotic anemones ( Figure S4 , lane 9).

The ARA was applied to determine if differences in nitrogenase activity were present in symbiotic versus aposymbiotic Aiptasia at ambient temperature (27°C). Over 24 h, conversion to ethylene was easily detected in fava-bean root nodules. Although acetylene was observed to decrease over 24 h in all Aiptasia samples, conversion to ethylene was not detected at 27°C ( Figure 9B ). The ARA was then repeated at 34°C. Root nodules showed a decrease in nitrogenase activity at elevated temperature, but was still easily detectable ( Figure 9B ). Both symbiotic and aposymbiotic Aiptasia produced similar but barely detectable amounts of ethylene (17.4 and 18.8 µM ethylene per gram anemone respectively, Figure 9B ). However, not all anemones remained alive after 24 h at 34°C.

Our study confirms prior reports that Aiptasia maintains core members of two key families, Alteromonadaceae and Rhodobacteraceae, as the most abundant across both symbiotic and aposymbiotic anemones, and at ambient and elevated temperatures. This is consistent with findings from Röthig et al. (2016) and suggests that these core members may not only be stable across multiple rearing conditions and research laboratories, but also might be specific to this strain of Aiptasia, as both studies use the CC7 clonal line. Host specificity, as opposed to environmental conditions, was once suggested not to be a major factor in bacterial community structure (Brown et al., 2017), but distinct microbiomes between the CC7 and H2 clonal lines of Aiptasia indicate otherwise (Herrera et al., 2017). Our study helps clarify some of these previously contrasting interpretations, affirming that stable members of two predominant bacterial families exist in Aiptasia and perhaps even within a specific clonal line.

However, the bacterial communities also varied dramatically across individual anemones, suggesting either tank effects, as observed previously (Hartman et al., 2020a; Xiang et al., 2022), or a stochastic element in microbiome acquisition and development, as supported by recent transplant studies (Damjanovic et al., 2017; Ziegler et al., 2017). Because of this variability, we focus primarily on the time-series experiment, which merged 15 anemones for each time point sample. Although lacking replication due to limitations of funds at the time of this experiment, multiple quality-control assessments of the DNA extractions were conducted to confidently pool anemones and sample multiple time points (see Materials and Methods). The strength of this experiment is the added capacity to observe trends in both prokaryotic (bacterial and archaeal) and eukaryotic microbial communities over time following temperature stress, with a much greater yield of bacterial OTUs and more consistent results (i.e., less variability between individual anemones or tanks).

Beyond Alteromonadaceae and Rhodobacteraceae, bacterial community structure appears rather flexible. Other bacterial families show considerable variation: Whereas Röthig et al. (2016) identified 14 other families across all their samples, we found 20 other families in ours. Although most families occurred in both studies, albeit at different levels of abundance, our control Aiptasia (symbiotic, 27°C) have microbiomes dominated by Pelobacteraceae ( Figure 4A ), represented in the community by the singular genus Pelobacter, a group not detected by Röthig et al. (2016). This family of Deltaproteobacteria, belonging to the order Desulfuromonadales, are fermenting anaerobes, and also have the ability to fix nitrogen (Schink, 1992; Akob et al., 2017). They were absent from aposymbiotic Aiptasia, suggesting that nitrogen fixation is more likely to occur when symbiosis with the algae is intact. This implication is in line with the findings of Xiang et al. (2022) in which diazotrophs and denitrifying bacteria were abundantly associated with symbiotic Aiptasia. In contrast, the family Kordiimonadaceae, represented by the singular genus Kordiimonas, showed the opposite pattern, being the second-most abundant family in aposymbiotic samples while absent in symbiotic anemones ( Figure 4A ). These bacteria are aerobic polycyclic aromatic hydrocarbon (PAH)-degrading bacteria found in a variety of shallow-water environments (Kim and Kwon, 2010; Math et al., 2012). The strong associations of Pelobacter and Kordiimonas with different symbiotic states are not described in any prior research on Aiptasia’s microbiome. Hence, caution must be taken in microbiome studies as we consider variability in the bacterial communities of various Aiptasia cultures maintained long-term in different laboratories conducting these studies, as well as the impacts of producing aposymbiotic states (e.g., use of the photosystem-II inhibitor DCMU). Interestingly, these dramatic microbiome shifts due to symbiotic state did not extend to surrounding seawater communities, which were quite similar ( Figure 4B ), but with large contributions from families not found in the anemones (e.g., Sphingomonadaceae, Chitinophagaceae) or in prior studies, though data are sparse.

Despite the different microbiomes present in symbiotic versus aposymbiotic animals, we observed clear trends in bacterial-community responses to heat stress corresponding to the period of bleaching in symbiotic anemones ( Figures 5 , 7 ). The dominant Pelobacteraceae in symbiotic Aiptasia dropped dramatically following heat stress, recovering only after bleaching on days 5-7 ( Figure 7A ), with several other families also being more (e.g., Saprospiraceae) or less (e.g., Alteromonadaceae) represented. Previous research shows coral- and Aiptasia-associated microbiomes exhibit higher diversity while undergoing bleaching caused by heat stress (Zaneveld et al., 2017; Ahmed et al., 2019). This change in diversity has been attributed to the Anna Karenina Principle (AKP), in which animals undergoing dysbiosis possess higher variation within microbial community composition compared to healthy animals (Zaneveld et al., 2017), attributed to the environmental stress limiting the ability of the host or its microbiome to regulate community composition. The outcome of a multi-year heat-stress experiment in Aiptasia (Ahmed et al., 2019) was in accordance with the AKP, but in our short-term heat-stress study, alpha-diversity indices during bleaching of symbiotic anemones were largely unchanged under heat stress ( Table 2 ).

However, we did observe increased abundance of rare taxa during the combined treatments of (i) lack of algal symbionts and (ii) animals and seawater under elevated temperatures ( Figures 7A, B ). Therefore, if we interpret aposymbiotic Aiptasia to be analogous to bleached corals and potentially experiencing some stress in this state, our results suggest that the AKP may explain the combined effects of multiple short-term stressors (this study) or a sustained single stressor (Ahmed et al., 2019). In aposymbiotic Aiptasia, the most notable change was a larger proportion of rare community members, which came to dominate the heat-stressed anemones by days 5-7 ( Figure 7B ). Among these were potential pathogens including Vibrio species, which exhibited a steady increase in aposymbiotic animals undergoing heat stress, as well as their surrounding seawater ( Figure 10A ), but not in symbiotic Aiptasia or any non-heat-stressed animals ( Figure 10B ). Dominant Vibrio species identified included V. harveyi, V. parahemolyticus, and V. alginolyticus. In tropical corals, V. harveyi is a known causative agent of white band disease, resulting in a general loss of tissue and algal pigment (Luna et al., 2010). Similar necrotizing pathogens V. coralliialyticus and V. shiloi, originally isolated from tropical corals and inoculated into Aiptasia, also cause disease (Zaragoza et al., 2014). As we identified V. harveyi exclusively in aposymbiotic samples, this suggests that the dual stressors (i.e., lack of algae and elevated temperature) are synergistic. Microbiome-transplant studies have shown that aposymbiotic Aiptasia have a ~10X lower carrying capacity than symbiotic Aiptasia (Costa et al., 2021), and perhaps the combination of thermal stress and lower microbiome occupancy can enhance the success of pathogens in the holobiont. Follow-up studies using a growth-inhibition assay with cultured V. harveyi in the presence of Symbiodiniaceae could isolate a potentially interesting algal-bacterial interaction within the holobiont.

Archaea have seldom been described in the coral holobiont, and never in Aiptasia (Puntin et al., 2022). Crenarchaeota and euryarchaeota have been found in close association with the tropical coral Porites asteroides, accounting for upwards of 50% of their prokaryotic community (Wegley et al., 2004). Archaea may have a role in methanol oxidation and ammonia oxidation (Siboni et al., 2008). In this study, we observed euryarchaeota present in the seawater, but not polyps, of heat-stressed aposymbiotic Aiptasia, although low sequence homology (87%) precluded identification. The presence of archaea exclusively in the environment but not in anemones themselves aligns with observations of archaea’s tendency to reside exclusively in coral mucus but not corals themselves (Frade et al., 2016). Further studies distinguishing anemone mucus from the seawater environment are warranted to understand the role of these novel archaea associated with the Aiptasia holobiont.

Since Pelobacter, a potential nitrogen fixer, was specifically associated with symbiotic anemones, and more present following heat stress at 34°C, we decided to further investigate the role of nitrogen fixation in Aiptasia. But despite using PCR primers that should easily amplify Pelobacter nifH ( Figure S3 ), we found, if anything, more evidence for genomic nifH in aposymbiotic Aiptasia ( Figure S4 ), but little evidence of actual nifH expression ( Figure 9A ) or nitrogen fixation, as measured by the ARA ( Figure 9B ), for both symbiotic states. Our results are vastly different from what has been observed in autotrophic coral species that display both nifH expression and nitrogenase activity (Pogoreutz et al., 2017), but this difference may also be attributed to the different nifH primers used between the two studies. Interestingly, our findings are more consistent with the highly heterotrophic coral species that Pogoreutz et al. (2017) examined, in which nifH abundance was low and nitrogen-fixation activity was also below the detectable limit. Thus, it would appear that cultured laboratory strains of Aiptasia may rely on heterotrophy when nitrogen is limiting. Furthermore, our observations of possible increased nitrogen fixation in heat-stressed Aiptasia ( Figure 9B ) align with observations of increased nitrogen-fixation activity under heat stress in certain corals (Rädecker et al., 2015; Cardini et al., 2016; Bednarz et al., 2019; Rädecker et al., 2021). But the lack of replication and obvious death of anemones during the 24-h heat stress in the ARA makes this observation very tenuous.

As Aiptasia laboratory cultures have been maintained for more than 30 years (Weis et al., 2008), it is possible that lack of diazotrophy is an adaptation to a more heterotrophic lifestyle, and its reputation as a notorious fast-growing contaminant of marine aquaria also suggests a different trophic strategy. Wild strains should also be examined for nitrogen-fixing microbiomes, but our results emphasize using caution when extrapolating results from laboratory strains of Aiptasia to coral-reef ecosystems.