For all experiments, C. xamachana medusae were collected by hand from two sites in Key Largo, FL, USA. The “Atlantic” site (25.086669, −80.453382) is near-shore and receives water flow from the Atlantic Ocean. The “Bay” site (25.079426, −80.453009) is located off docks leading into the Florida Bay ( Supplementary Figure 1 ). Water temperature was measured at collection sites using pendant HOBO^® data loggers (MX2201) were deployed from March 12 to September 9, 2022, at a depth of 1.5 m to reflect the depth at which C. xamachana are regularly observed. Daily mean temperatures and ranges (maximum - minimum) were calculated using 15-minute interval data ( Figures 1A, B ). Statistical differences between mean daily temperature and daily temperature were determined using t-tests in R with the stats package (v. 4.3.0; R Core Team, 2023).

In May 2021, approximately 50 C. xamachana medusae (28–95 mm) were collected from each of the two sites ( Supplementary Figure 1 ) and transported to Auburn University where they were acclimated for two weeks in a closed-sump aquarium system under ambient conditions (26 °C ± 1 °C; 36 ppt salinity). Animals were maintained on a 12:12 light:dark cycle and fed Artemia three times per week. Morphological characteristics (bell diameter, sex, and appendage morphology) were recorded, and animals were re-acclimated for one week to compensate for handling stress.

Acute heat stress trials were performed such that 10 animals were monitored simultaneously, and trials were repeated five times with different individuals (10 animals x 5 blocks = 50 total animals measured). Within each block, five individuals from each collection site were chosen at random. The evening before each trial, individual medusae were placed into 500 mL containers with mesh sides in the experimental tank (water conditions were identical to acclimation tank). Water temperature was controlled using two submersible rod heaters (Finnex Deluxe Titanium Tube Heater, 800W) connected to a digital temperature controller (SunTHIN WiFi Digital Temperature Controller 1250W, 10A). Experimental tank temperatures were monitored using pendant HOBO^® data loggers (MX2201) recording temperatures at 15-minute intervals. Heat stress trials were performed at the same time each day to reduce potential variability associated with photoperiod (Nath et al., 2017). Water temperature was increased 1°C/hour; 30 minutes after each temperature increase, 1-minute videos were recorded in triplicate using Blink security cameras (Amazon) suspended above the experimental tank. Trials were complete when all medusae reached lethality, defined here as either 0.5 x baseline pulsation rate or < 10 pulses/minute. This criterion corresponded with tissue disintegration and was often accompanied by a separation of arms from bell tissue (Toullec et al., 2024). Dead animals were immediately removed from the experimental tank. Lethal temperatures and bell diameters at time of death were recorded for each individual.

In May 2022, approximately 50 C. xamachana medusae of each color morph (blue and uncolored/brown) were collected from the Atlantic site ( Supplementary Figure 1 ) and transported to Auburn University where they were randomly placed into two sump systems, each with 4 individual tanks, and maintained as described above. After a five-day acclimation period, morphological characteristics (bell diameter, sex, and appendage morphology) were recorded. Individuals were photographed at a standard height (0.5 m) using a Canon Rebel X camera with a size and color standard. Tissue samples were collected from arm tips to measure symbiont density (stored at 4°C) and isolate genomic DNA (flash-frozen in liquid nitrogen and stored at −80°C). Animals were acclimated for three days to compensate for handling stress.

The temperature in one of the two sump systems was increased 2°C per day and then maintained at 33°C (± 1 °C) for 30 days (N =43; these animals are referred to as “elevated”). Individuals in the remaining tanks (N = 42) were maintained in ambient conditions (26 °C ± 1°C). Temperatures were monitored using pendant HOBO^® data loggers (MX2201). After the acclimation period, bell diameters were measured, and arm tissue was sampled to quantify symbiont density (stored at 4°C).

Acute heat stress trials were performed on medusae three days after post-acclimation measurements. Trials were performed as described in Experiment 1, with the additional parameters that 1) animals of each color morph were evenly distributed among the blocks; and 2) “elevated” animals were transferred to the experimental tank when the temperature matched that of their acclimation tank (33 °C).

Rates of bell pulsation were quantified using the videos captured during acute heat stress experiments. Videos were viewed once per individual within the camera view. Pulsation rate was calculated as the number of pulses per minute. Bell pulsation rates vary uniquely among Cassiopea sp. individuals (Rowe et al., 2022). To account for this variation and assess the influence of heat stress on bell pulsation, baseline pulsation rates were calculated in each individual in the absence of any experimental manipulation. For all experiments in 2021 and the medusae that were acclimated at ambient temperatures in 2022, baseline rates were calculated after the animals were transferred to the experimental tanks and allowed to acclimate overnight to account for handling stress. In medusae acclimated at elevated temperatures (2022 trials), baseline rates were calculated in the experimental tank at 32°C. For all data shown, the individual baseline pulsation rate was subtracted from the pulsation at each temperature. All raw data, including the baseline pulsation rates and un-adjusted rates are available at https://github.com/mem0294/Maloneyetal2024.

To test the effects of collection site or acclimation status, appendage color, and temperature on unadjusted bell pulsation rates, a generalized linear mixed-effect model (GLMM) with a Poisson link-log function was conducted in R using the package lme4 (v.1.1-7, Bates et al., 2015). In these designs, the location or acclimation state, appendage color, and temperature during heat stress were fixed effects, and individuals and the date of the experiment (block) were coded as random effects. To test the effects of location or acclimation, appendage color, and heat stress temperature on the change in bell pulsation rates from baseline rates, a linear mixed effect model (LMM) was conducted with the R package lme4 (v.1.1-7, Bates et al., 2015). A LMM was used to examine differences in lethal temperature between location/acclimation and appendage colors. In this design, location/acclimation and appendage color were fixed effects; the date of the heat stress trial (block) was a random effect.

For all analyses, assumptions of normality and homogeneity of variance were graphically assessed using histograms, residual plots, and quantile-quantile plots. When there was a significant main effect (p < 0.05), Tukey’s post hoc mean comparisons were conducted using the multcomp package (v.1.4-25, Hothorn et al., 2008).

Images taken before the acclimation treatment were analyzed in Adobe Photoshop CC to quantify number of pigmented appendages, color information, and blue coverage (percent). The number of pigmented appendages was quantified using FIJI (Schindelin et al., 2012) using the “Cell Counter” plugin (O’Brien et al., 2016). Appendages were counted if they met any of the following criteria: 1) contained a color dissimilar to that of the bell or frilled sections (this includes but is not limited to: blue, green, yellow, purple) or 2) morphologically matched that of a typical “large vesical” (as described in Hummelinck, 1968). The presence of “pigmented” appendages was not used to distinguish blue vs. brown individuals.

RGB (red, blue, and green) and HSB (hue, saturation, brightness) values were quantified using the Color Sampler tool. Hue refers to true color (e.g., red, green, or blue) and is measured as degrees on a color wheel, such that 0° (and 360°) is red. Saturation is the intensity or richness of color and is measured by a percentage, such that 0% indicates “no color” and 100% indicates “intense color”. Brightness is also measured as a percentage such that 0–100% corresponds to black–white. Four measurements were recorded for each individual: 1) a pigmented vesicle (if present); 2) a non-pigmented oral arm appendage; 3) the bell tissue; and 4) the blue color standard that was included in each image and served as a control for variation in imaging. To normalize across images, the HSB values of the blue color standard were subtracted from each of the jellyfish measurements. A Pearson correlation was conducted in R to assess the relationship between lethal temperature and adjusted saturation percent.

The coverage of blue appendages present in individual medusae was quantified using the number of pixels (in percent) associated with blue in the R package countcolors (v.0.9.1, Weller and Westneat, 2019). Image backgrounds were removed in Adobe Photoshop and replaced with white; images were manually edited to ensure that the entire medusa was included. The countcolors package quantifies all pixels in an image according to RGB values translated to a 0–1 scale, rather than the traditional 0–255 (Weller and Westneat, 2019). The package was run as follows: ignore white (RGB [1, 1, 1]) to exclude the background pixels and identify the “blue” pixels (RGB [0.33, 0.30, 0.25] with an extended radius of 0.1. This color range was determined by quantifying RGB values for pigmented appendages sampled across all blue jellyfish and averaging the value. Note that, because these organisms are not bleached, the background tissue color is light brown due to the presence of the algal endosymbionts. Therefore, the color values of the pigmented appendages fall within a brownish gray, rather than pure blue. The countcolors package subsequently exports a new image in which pixels within the range are masked as magenta (RGB [1, 0, 1]). The accuracy of the masked images was verified manually and adjusted as necessary. Finally, countcolors was used on the masked images to calculate the percentage of magenta pixels in the image. A Pearson correlation was conducted in R to assess the relationship between lethal temperature and amount of blue coloration.

Medusa arm tip pieces were added to pre-weighed 1.5 ml centrifuge tubes and weighed to determine wet mass. Frozen tissue was homogenized in 500-700 μL of artificial seawater (Instant Ocean) using a disposable pestle and an electric homogenizer. To separate host tissue from the Symbiodiniaceae cells, homogenized samples were centrifuged at 800 x g for 10 minutes. Algal pellets were washed in 500 µl artificial seawater, vortexed to mix and re-homogenized as above. Samples were washed three times in artificial seawater (400 x g for 10 minutes) and resuspended in 150 µl Z-Fix (ANATECH). Symbionts were counted using an Improved Neubauer hemocytometer and normalized to the wet mass to calculate density. Eight counts were performed for each individual. A repeated measures ANOVA in R was used to examine symbiont density before and acclimation. In this design, acclimation treatment and appendage color were categorized as fixed effects; individual medusa ID was used as the repeated measure, and therefore the random effect.

DNA from arm tissue was extracted using the E.Z.N.A. Tissue DNA Kit (Omega Bio-Tek) according to manufacturer’s instructions. DNA concentrations were quantified using the dsDNA Broad-Range Qubit assay (Invitrogen). The mitochondrial gene cytochrome c oxidase subunit I [COI] was amplified using custom-designed primers ( Supplementary Table 1 ; Ye et al., 2012). Thermal cycling conditions were 3 min at 95°C for initial denaturation, followed by 35 amplification cycles (denaturation at 95°C for 35 s, annealing at 49°C for 40 s and extension at 72°C for 50 s) and a final extension for 7 min at 72°C. Amplified products were cleaned (Zymo Research Genomic DNA Clean and Concentrator-10 kit) and quantified using the dsDNA Broad-Range Qubit assay (Invitrogen). Purified products were sequenced using Sanger chemistry with custom primer MM1 ( Supplementary Table 1 ). COI sequences were aligned to publicly available sequences from C. xamachana (Genbank accession NC_016466.1; Ohdera et al., 2019) and C. andromeda (OP503353.1 and OP503325.1; Kayal et al., 2012) using Geneious v. 2022.2.2.

Medusae collected from the Bay (n = 6) and Atlantic sites (n= 5) were used to characterize symbiont communities. Tissues samples were collected from arms, stored in ethanol and used for DNA extractions as described. ITS-2 regions were amplified using GoTaq MasterMix (Promega), and Symbiodiniaceae-specific primers (final concentration 400 nM) with adaptors for Illumina sequencing ( Supplementary Table 1 ). Reactions were amplified as follows: 95°C for 3 min, 25 cycles of 95°C for 30 sec, 55°C for 30 sec, 72°C for 30 sec, and 72°C for 5 min. Samples were purified with the Genomic DNA Clean and Concentrator-10 kit (Zymo Research) and sequenced on the Illumina MiSeq platform (150 bp paired-end reads; Georgia Genomics).

The R package dada2 (Callahan et al., 2016) was used to identify and quantify the Symbiodiniaceae taxa present in each sample. After removing chimeric sequences, valid amplicon sequence variants (ASVs) were identified and enumerated. To identify the proportion and identity of the symbiont community in each sample, we created a BLAST database containing annotated ITS2 sequences from the SymPortal framework (https://symportal.org).

Given the morphological variation and potential for cryptic species among Cassiopea sp., (Muffett and Miglietta, 2023), the species identity of collected animals was confirmed. DNA isolated from medusae used for the 2022 acute heat stress experiment was used to amplify the COI gene for sequencing. All 30 sequences exhibited >99.6% sequence identity with the Cassiopea xamachana COI sequence (Kayal et al., 2012; Supplementary Figure 2A ). Further, 11 representative animals from the collection sites were selected for ITS-2 symbiont typing. These medusae were overwhelmingly dominated by Symbiodinium A1 ( Supplementary Figure 2B ), which is consistent with studies showing strong host-symbiont fidelity between C. xamachana and S. microadriaticum (LaJeunesse, 2017). Thus, despite the observed morphological variation among individuals, medusae used in these studies are C. xamachana.

To determine if environmental history affects how C. xamachana responds to heat stress, medusae were collected from two sites that vary significantly in daily temperature range (“Atlantic” and “Bay” sites; Supplementary Figure 1 ). Temperature data collected from the two sites revealed that, although these locations had similar average daily temperatures (30.1°C in the Bay and 29.9°C in the Atlantic; Figure 1A ), the Atlantic site water temperatures were significantly more variable than those in the Bay site (p_t-test < 0.001, Figure 1B ). On average, water temperatures at the Atlantic site varied 4.55°C per day compared to 1.73°C at the Bay site. This variability may reflect changes in water depth due to tidal influx.

Animals from each site were exposed to an acute lethal heat stress (starting from ambient temperature and increasing 1°C/hour until lethality). Bell pulsation rates were used to monitor organismal response. Notably, no significant differences were observed in the change in bell pulsation rates among animals collected from the two sites ( Figure 1C ; Supplementary Figure 3 ; Supplementary Table 2 ). The rates of bell pulsation increased ~3.1 pulses per minute for each 1°C increase in water temperature (p_GLMM < 0.001, Supplementary Table 3 ) until 37°C, at which point pulsation rates dropped dramatically just prior to mortality. Post-hoc analyses identified statistical differences in the change in pulsation rates at several temperatures relative to baseline ( Supplementary Table 4 ). Additionally, no significant difference was observed in average lethal temperatures of C. xamachana collected from the Atlantic (39.81°C), and Bay sites (39.52°C; p_LMM = 0.356, Figure 1D , Supplementary Table 5 ). Notably, despite surviving to 40°C, we observed no evidence of symbiont loss (bleaching) in medusa from either collection site. Together, these data indicate that, despite the difference in daily temperature range and variability between collection sites, C. xamachana from both locations displayed similar physiological phenotypes in response to acute heat stress.

To determine if C. xamachana can acclimate to elevated temperatures, medusae (N=100) were collected from the Atlantic site and maintained at either ambient (26°C) or elevated (32°C) temperature for 30 days ( Supplementary Figure 4A ). The different acclimation temperatures did not cause significant differences in bell diameter or symbiont density ( Supplementary Figure 4B, C ). In fact, medusae in both treatment conditions exhibited reduced bell diameters and increased symbiont density ( Supplementary Figure 4B, C ), which may reflect the transition to the artificial aquarium system. Following acclimation, 40 individuals (N_ambient = 20, N_elevated = 20) were randomly selected for heat stress trials.

Following the 30-day acclimation period and after the first group of animals was subjected to experimental heat stress, an unexpected mechanical failure in the ambient tank (26°C) caused a temperature decrease of 5°C for 24 hours. The remaining individuals were thus treated as “cold-shocked”. To account for the possibility that this affected subsequent responses, shock status was added as a fixed effect for all models. Notably, no significant interaction of shock status was observed with any other treatment factors. We have therefore interpreted fixed effects (temperature, acclimation treatment, and appendage color) equally.

Animals were exposed to an acute heat stress starting from the acclimation temperature and increasing 1°C/hour until lethality. As in the initial study, rates of bell pulsation steadily increased with temperature until animals approached lethality ( Figure 2 ; Supplementary Figure 5 ). Acclimation temperature had a significant effect on the change of bell pulsation rates from the start of heat stress (p_LMM = 0.008, Figure 2A , Supplementary Table 6 ). Notably, medusae acclimated in elevated conditions survived to higher temperatures compared to those acclimated at ambient temperatures. The average lethal temperature for C. xamachana acclimated at ambient temperature was 39.97°C compared with animals acclimated at elevated temperatures (41.2°C, p_LMM < 0.0001; Figure 2B ; Supplementary Table 8 ). This data show that C. xamachana can acclimate to a consistent period of elevated temperatures and that this acclimation is beneficial in the context of acute heat stress.

Given the lack of significant differences in medusae collected from two sites ( Figure 1 ), we were interested in identifying other factors that might influence C. xamachana physiology during heat stress. We therefore investigated if the variation in color phenotypes is associated with survival during heat stress. Using the data from both experiments, C. xamachana individuals were categorized as either “blue” or “brown” based on the predominant color of their oral appendages and all data were re-analyzed in the context of appendage color ( Figure 3 ). Importantly, no relationship was observed between appendage color and either bell diameter or symbiont density ( Supplementary Figure 6 ). The change in rate of bell pulsation during the acute heat stress experiments was not significantly different between medusae with blue appendages or brown appendages in either experimental year (2021 experiment, p_LMM = 0.699, Figure 3A ; 2022 experiment, p_LMM = 0.681, Figure 3B ). However, in both experiments, medusae with blue appendages survived significantly higher temperatures than those without ( Figure 3C ; 2021 experiment, p_LMM = 0.003, Supplementary Table 5 ; 2022 experiment, p_LMM < 0.007, Supplementary Table 8 ). This effect was also observed in medusae collected from both sites in 2021, such that individuals with blue appendages from the Atlantic site survived to significantly higher temperatures than those from the Bay site (p_LMM = 0.0013; Figure 3D ; Supplementary Table 9 ). In both blue and brown individuals, elevated acclimation treatment was significantly associated with higher survival (p_LMER < 0.0001; Figure 3D ).

To further investigate if there is a relationship between the amount of blue coloration coverage/saturation of pigmented appendages and lethal temperature during an acute heat stress, appendage morphology was quantified as the number of pixels within the average RGB of blue appendages. A significant, positive correlation was observed between the percentage of blue coverage within a medusa and lethal temperature ( Figure 4A , R = 0.39; p = 0.032). Similarly, we found a significant, positive correlation between the adjusted saturation of the appendage and lethal temperature ( Figure 4B , R = 0.41, p = 0.024). We however found no significant relationship between the number of pigmented appendages and lethal temperature ( Figure 4C , R = 0.26, p = 0.15). Analysis of the acclimation treatment data revealed positive correlations between percent blue coloration and lethal temperature in medusae acclimated in either temperature treatment, although this effect was stronger in ambient conditions (R = 0.55, p = 0.034), relative to elevated (R = 0.36, p = 0.19, Supplementary Figure 7A ). Positive correlations were also present for adjusted saturation ( Supplementary Figure 7B ). Representative medusa and their corresponding results for blue coverage, adjusted saturation, number of appendages, RGB and HSB are shown in Figure 4D .

Among symbiotic marine cnidarians, Cassiopea appears to be uniquely capable of thriving in a range of environmental conditions (Purcell, 2012; Toullec et al., 2024). Specifically, C. andromeda exhibits exceptional tolerance to repeated, intense (26–34°C) heat stress episodes (Banha et al., 2020; Béziat and Kunzmann, 2022). Here, using rates of bell pulsation as a measure of organismal physiology, we find that C. xamachana maintain homeostatic bell pulsation over a broad, but ecologically-relevant, temperature range (26–34°C). This contrasts with several species of tropical cnidarians in which bell pulsation rates change dramatically beyond a narrow temperature range (Béziat and Kunzmann, 2022; Gatz et al., 1973). In Cassiopea sp., bell pulsations create water currents that facilitate food capture (Gohar and Eisawy, 1960; Hamlet et al., 2011; Klein et al., 2016) and oxygenate the surrounding water (Rowe et al., 2022). In the context of temperature stress, increased rates of bell pulsation may serve as a mechanism to enhance metabolic processes during stress (Fitt et al., 2021; Mangum et al., 1972; McClendon, 1917). This phenomenon has been observed in other marine invertebrates responding to acute temperature stresses (e.g., crayfish, Ern et al., 2015; gastropods, Valles-Regino et al., 2022). In response to temperature fluctuations, Cassiopea sp. exhibit increased superoxide dismutase (SOD) activity (McCord and Fridovich, 1969), signs of oxidative stress (Regoli and Giuliani, 2014), and altered metabolic rates (Aljbour et al., 2019; Fridovich, 1973; Lesser, 2006; McCord and Fridovich, 1969; Regoli and Giuliani, 2014). Stress conditions, such as increased temperature beyond physiological limits, may increase the demand for metabolic energy and, in turn, increase reactive oxygen species (ROS) (Boveris and Chance, 1973; Speakman and Selman, 2011). This ability to overcome oxidative stress may underlie C. xamachana resilience to thermal stress. Future work should aim to clarify the physiological and metabolic benefits or costs of increasing bell pulsation and how this contributes to the remarkable thermotolerance of C. xamachana.

It has been proposed that high environmental variability can result in increased phenotypic plasticity (Lande, 2014). This theory is supported by several empirical studies in marine invertebrates, demonstrating that organisms living in fluctuating environments (e.g., drastic fluctuations in temperature or salinity) exhibit increased resistance to temperature stress associated with climate change (Burton et al., 2022; Godefroid et al., 2023; Kapsenberg and Cyronak, 2019; Rivest et al., 2017; Safaie et al., 2018). In contrast, here we demonstrate that C. xamachana collected from a site with high daily fluctuations in temperature exhibit the same physiological (bell pulsation rate) and survival (lethal temperature) response during heat stress than individuals collected from a site with stable daily temperatures. Similar experiments in other cnidarian species have produced contradicting results. For example, temperature fluctuations on daily or tidal timescales were suggested to be sufficient to promote thermal tolerance of some corals via acclimation, while high temperatures were short enough to avoid lethality (Castillo and Helmuth, 2005; Oliver and Palumbi, 2011; Safaie et al., 2018). However, the effect of temperature variability on heat tolerance is likely species-specific (Putnam and Edmunds, 2011; Schoepf et al., 2022; Voolstra et al., 2020).

Here we show that C. xamanchana are robust to thermal stress, and that this was not likely explained by inhabiting environments with variable daily temperatures. Several potential factors may contribute to this result. First, the temperature fluctuations between the two sites may not be sufficiently different to elicit distinct physiological effects. Corals from habitats such as the back-reef pools in American Samoa experience daily temperature fluctuations up to 5.6°C, even reaching daily extremes of >35°C (Barshis et al., 2010); whereas the nearby, less-variable forereef experiences seasonal maximum daily temperature fluctuations of 1.8°C (Barshis et al., 2018; Craig et al., 2001). These corals have phenotypic differences consistent with local adaptation of thermal performance (Palumbi et al., 2014). Similarly, either the temperature or the duration of daily maximum temperature in the Atlantic site may be insufficient to induce increased thermal tolerance (either via acclimation or local adaptation). While our COI data revealed that all animals used in this study are C. xamanchana (consistent with Muffett and Miglietta, 2023), these data are unable to resolve population-level differences or assess local adaptation. Thus, it is possible that C. xamachana sampled from both sites represent the same population and are thus robust to stable and fluctuating temperatures regardless of the site they are found. These results are consistent with the overall observed robustness of C. xamachana relative to other marine invertebrates (Aljbour et al., 2019; Banha et al., 2020; Béziat and Kunzmann, 2022; Fitt et al., 2021; McGill and Pomory, 2008; Medina et al., 2021; Ohdera et al., 2018; Rowe et al., 2022). In fact, Cassiopea sp. thrive in human-disturbed environments (Stoner et al., 2011, 2016) and are expanding their habitat range. Our results may therefore reflect the underlying thermal tolerance of C. xamachana rather than environmental conditions of specific collection sites.

As oceans warm for more prolonged periods due to global change, marine organisms must adjust their physiology to tolerate greater extremes (Coumou and Rahmstorf, 2012; Lopez et al., 2018; Morley et al., 2019). Short-term temperature stress can lead to subsequent increases in thermal tolerance in cnidarians (Baker et al., 2008; Kenkel and Matz, 2016; Kirk et al., 2018). The work presented here indicates that acclimation to a warmer environment may promote survival under acute heat stress and thus may prepare C. xamachana medusae for future warming conditions. Given the rapid heating of shallow water environments, C. xamachana may have an advantage if previous warming allows them to acclimate; it is possible that the previously encountered temperature “extreme” in our acclimation experiment could have buffered the effect of the acute heat stress, thereby allowing them to survive to higher lethal temperatures (i.e., stress memory; Crisp et al., 2016; Li et al., 2020).

These results, while novel for C. xamachana., are not uncommon among marine ectotherms. In natural experiments, corals of the Great Barrier Reef pre-exposed to thermal stress exhibited protection from subsequent heat waves (Ainsworth et al., 2016; DeMerlis et al., 2022). It is hypothesized, that an organismal thermal threshold might be a plastic trait, in which some individuals are able to extend the upper thermal limit over time (Buckley and Huey, 2016; Leung et al., 2021). These observations are consistent with the laboratory-based experiments presented here and by others investigating cnidarian thermal tolerance (Ainsworth et al., 2016; Bay and Palumbi, 2015; Bellantuono et al., 2012a, 2012b; Middlebrook et al., 2008). Cassiopea appear to be resilient to temperature stress, which may be the result of a hardening capacity, or the ability to increase thermotolerance due to experiencing a prior non-lethal temperature stress (Sgrò et al., 2010). The results presented here emphasize Cassiopea’s ability to acclimate to warming temperatures but may also predict their invasion potential and survival in more diverse environments. Some studies have already investigated Cassiopea’s invasion into more temperate environments (Rowe et al., 2022), however it is possible, due to their ability to acclimate, they may expand their range into warmer habitats.

How temperature-dependent processes differ among color morphs could impact species survival and influence how polymorphisms evolve and/or persist (Thompson et al., 2023). In the present study, we observed differences in C. xamachana response to heat stress in medusae containing blue appendages compared to those lacking blue appendages. This finding was consistent across two years of experimentation using individuals collected from different sites, which demonstrates the strength of our hypothesis that pigmented appendages may benefit C. xamachana during extreme heating events. This is the first record of appendage color contributing to fitness traits in C. xamachana. While blue appendage color is common among some Cassiopea species, other color morphs are also common, including red, blue, purple, green, brown, and white (Lampert et al., 2012). Cassiopea sp. also vary dramatically in the level of pigmentation, appendage shape and appendage number (Lampert et al., 2012; Rowe et al., 2022). Color has been associated with camouflage, sexual selection, and social interactions in many organisms (Cuthill et al., 2017), although these roles are unlikely in Cassiopea. Given their habitat (shallow, clear seawater), and ecology (epibenthic, with brown coloration mimicking the substrate), bright blue coloration makes individuals more obvious. Additionally, Cassiopea males freely release sperm into the seawater and females brood larvae and therefore do not undergo coloration-based mate selection. Based on the results presented here, we hypothesize that the Cassio Blue protein provides Cassiopea with an advantage in thermal stress conditions at the level of cellular physiology.

Several hypotheses on the function of Cassio Blue have been proposed. One possibility is that Cassio Blue facilitates optimal photosynthesis via mitigation of solar radiation (Phelan et al., 2006), although Cassio Blue pigmentation does not strongly co-localize in tissue with photosymbionts. Other chromoproteins have been proposed to absorb excess light and thus optimize the environment for algal symbionts (Ferreira et al., 2023; Salih et al., 2000; Smith et al., 2013). It has also been hypothesized that the pigmented areas in Cassiopea provide a refuge for the symbiotic algae and that certain colors correlate with specific algal strains. However, data presented here and by others (Lampert et al., 2012) indicate that this is not the case. Different color morphs harbor identical symbiont communities ( Supplementary Figure 2 ), and symbionts were present in both pigmented and non-pigmented tissues. Furthermore, an ortholog of Cassio Blue is present in the blue jellyfish, Rhizostoma pulmo, which does not harbor symbionts, making it likely that the primary role for this protein is related to host, rather than symbiont, physiology (Bulina et al., 2004; Lawley et al., 2021).

The Cassio Blue protein is composed of Frizzled and Kringle domains, neither of which have been previously associated with pigmentation or thermal stress responses (Lawley et al., 2021). However, based on physiological roles for fluorescent proteins in other cnidarians and chemical properties of Cassio Blue, we propose two hypotheses for how this protein may promote survival in heat stress. In Cassiopea sp., heat stress induces signs of oxidative stress (Aljbour et al., 2019) and increased production of reactive oxygen species (ROS). To attenuate potential damage from ROS, some cnidarians express proteins with antioxidant properties. For example, fluorescent proteins expressed in several species of scleractinian corals serve as scavengers for ROS (Palmer et al., 2008). As an alternate hypothesis, Cassio Blue is associated with transition metals, including Cu, Fe, and Zn (Phelan et al., 2006). It is therefore possible that this protein serves as an important reservoir for these trace metals, which are required for metabolic processes such as the protein synthesis and enzyme activity needed for physiological acclimation (Reich et al., 2021). Metals have also been implicated in several aspects of coral physiology, specifically relating to the ability of algal symbionts to photosynthesize (Ferrier-Pagès et al., 2018; Reich et al., 2021, 2023). Symbiodiniaceae have higher metal requirements to initiate electron transfer between photosystems (Camp et al., 2022; Raven et al., 1999; Reich et al., 2020; Rodriguez et al., 2016). These hypotheses are not mutually exclusive; Cassio Blue may be involved in both functions. Future work will be necessary to examine the mechanistic role of Cassio Blue in C. xamachana physiology.

Trade-offs between fitness benefits and energetic costs of producing pigmentation may impact organismal physiology (Calsbeek et al., 2010), however these factors may not be relevant during acute heat stress in Cassiopea. In some organisms, selection pressures for coloration may result in variation in color traits (Zajitschek et al., 2012). However, in C. xamachana, blue appendages do not appear to influence whole-animal performance traits, such as bell pulsation, but may instead contribute to overall survival through alternative mechanisms (Husak and Fox, 2008; Irschick et al., 2008; Zajitschek et al., 2012). Blue color in many organisms can be achieved by combining a chromophore with an apoprotein (Bulina et al., 2004), which is not a common phenomenon due to the complex structure of blue pigments. Although the chromophore for Cassio Blue has not yet been identified, it is likely that the benefits of blue pigmentation outweigh the potential costs of producing the protein.

Color and fluorescent proteins are extremely common in Cnidaria and have been associated with immunity functions, growth enhancement, and photoprotection (D’Angelo et al., 2012; Palmer et al., 2008, 2009; Salih et al., 1997; Smith et al., 2013). While functional studies on blue pigmentation in Cassiopea sp. have been somewhat limited, considerable insights can be gained from work in corals. Populations of the stony coral Montastraea cavernosa, which have green and red color morphs, possess the same gene, but exhibit differences in mRNA abundance between morphs (Ferreira et al., 2023; Kelmanson, 2003). According to a study by Gittins et al. (2015), copy number and expression level of the RFP gene was found to be an important factor in the intensity of red fluorescence in Acropora millepora (Ferreira et al., 2023). Like what has been studied in corals (Bollati et al., 2022; Dove, 2004; Lyndby et al., 2016; Salih et al., 2000), Cassiopea sp. appendage colors other than blue could generate different light microenvironments, which may be important for photoprotection in shallow waters (Lyndby et al., 2023).