The effects of dissolved organic matter supplements on the metabolism of corals under heat stress

Octocorals represent a major alternative group in the benthic community of reefs that have diverged from hexacoral dominance. Despite their phototrophic symbionts, supplementing their diet with heterotrophic sources may promote their growth, particularly when compared to hexacorals in bleaching conditions. However, limited comprehensive data exists on octocorals' trophic ecology, especially regarding their ability to feed on dissolved organic matter (DOM), which comprises the largest pool of organic matter in reefs. This study aims to investigate the ability of two octocorals (Sarcophyton glaucum and Lobophytum sp.) to feed on DOM and compare this ability to that of hexacorals, such as Stylophora pistillata and Turbinaria reniformis. The study measured the net fluxes of DOM under varying DOM concentrations and under heat stress. The results demonstrate that all coral species were net producers of DOM at ambient concentrations, but became net consumers once seawater was supplemented with DOM. Furthermore, our study highlights a relationship between coral physiological responses to heat stress and DOM uptake. Corals that increased (S. pistillata) or maintained (S. glaucum and Lobophytum sp.) their DOM uptake rates at high temperature were the most resilient to heat stress. In contrast, T. reniformis exhibited lower DOM uptake rates at high temperatures, which was associated with significant bleaching. Understanding the ability of corals to feed on DOM may, therefore, provide insight into the resilience of species under ocean warming conditions.

Coral reefs exhibit high primary productivity and species diversity despite thriving in nutrient-poor waters (Odum and Odum 1955;Crossland et al. 1991).The success of symbiotic corals (including hexacorals and octocorals), which are the main reef builders in these oligotrophic environments, is attributed to their ability to rely on both autotrophic and heterotrophic feeding, thus exploiting a wide variety of energy and food sources for reproduction, development, and growth (Houlbrèque and Ferrier-Pagès 2009;Ferrier-Pagès et al. 2021).Photosynthetic Symbiodiniaceae living in coral tissue are able to convert light energy and dissolved inorganic nutrients into organic compounds, such as amino acids, fatty acids, and sugars, which are transferred to the host for their own metabolic demands (autotrophic nutrition, Muscatine et al. 1981;Tremblay et al. 2012a).The host, in turn, can feed on a variety of organic food sources, from dissolved organic matter (DOM) to pico-and nanoplankton (e.g., bacterioplankton) and even predate on larger zooplankton (heterotrophic nutrition, Houlbrèque and Ferrier-Pagès 2009).
Present-day corals suffer from increasing heat waves that disrupt the symbiotic association with their Symbiodiniaceae, causing so-called bleaching events that deprive the coral host of autotrophic nutrient supply (Rädecker et al. 2021) and lead to (sometimes major) coral death.Therefore, corals that are able to increase their heterotrophic capacity to compensate for the loss of autotrophy have been identified as potential winners under future reef conditions (Grottoli et al. 2006;Ferrier-Pagès et al. 2018).Understanding coral mixotrophic nutrition is thus essential to predict their resilience under heat stress.However, the heterotrophic uptake of nutrients is highly species-specific, and depends on the type of food available to the corals (Grottoli et al. 2006).In particular, the importance of DOM for coral nutrition, under normal growth conditions or heat stress, has been poorly studied, although DOM is increasingly recognized as a significant food source for many benthic species, including corals (Tremblay et al. 2012b;de Goeij et al. 2013;Levas et al. 2016).DOM is, by far, the largest pool of organic matter on reefs (Atkinson and Falter 2003;de Goeij and van Duyl 2007;Tanaka et al. 2011).It consists of a large "recalcitrant" fraction that is not readily bioavailable, but the smaller bioavailable "labile" portion (up to 20-25%; de Goeij and van Duyl 2007) is still orders of magnitude larger than the concentration of particulate organic matter (POM) or even inorganic nutrients in reef waters (Atkinson and Falter 2003).In contrast to zooplankton, which generally occurs in low abundance in reef waters, except during nighttime (Yahel et al. 2005), DOM constitutes a continuously available food source and may, therefore, sustain corals during bleaching events.Potential sources of DOM in coral reefs are numerous, from algae excreting a high proportion of their photosynthates to "sloppy" feeding that results in DOM leaking from broken prey (Moller 2004;Wijgerde et al. 2011).Although healthy corals generally release DOM as part of mucus (Bythell and Wild 2011;Hadaidi et al. 2019), some corals can also take up DOM in the form of urea and free amino acids (Hoegh-Guldberg and Williamson 1999;Grover et al. 2006Grover et al. , 2008;;Gori et al. 2014).Under stress conditions, including bleaching, corals have been shown to reverse from net release to net uptake of DOM to mitigate or compensate for the loss of autotrophic nutrients (Tremblay et al. 2012b;Levas et al. 2016).
With the exception of one study on octocorals (Bednarz et al. 2012), uptake and release of DOM have only been measured in a few hexacoral species of the order Scleractinia, although they are not the predominant species in eutrophied, DOM-enriched reefs (Baum et al. 2016).Indeed, high abundances of octocorals have been reported in eutrophied reefs of the Indo-Pacific and Caribbean (Norström et al. 2009;Inoue et al. 2013;Baum et al. 2016), representing the second most abundant macrobenthic group in reefs (Fabricius and Alderslade 2001).Octocorals are thus found to be the most abundant alternative community on reefs where hexacoral cover declined over the last decades due to reduced water quality, increased sea surface temperatures, or pollution (Reverter et al. 2022).Octocorals can quickly colonize reefs due to their high fecundity and fast growth rates, and these opportunistic life history features may be facilitated by their better resistance to DOM enrichment (Vollstedt et al. 2020) or heat stress (Steinberg et al. 2022) than hexacorals.They also appear to be more heterotrophic, as they have lower rates of photosynthesis and uptake of dissolved inorganic nutrients than hexacorals (Pupier et al. 2019(Pupier et al. , 2021)), but much higher grazing rates on phyto-and zooplankton (Fabricius et al. 1995;Fabricius and Klumpp 1995;Fabricius and Dommisse 2000;Rossi et al. 2020).They may also largely feed on DOM, although little comprehensive data exists on this trophic ecology aspect (Bednarz et al. 2012).Therefore, further comparative studies of the trophic ecology and stress response of octocorals and hexacorals are needed to predict which factors promote the growth of octocorals under current and future environmental conditions.
In this study, we tested the bleaching response of two species of octocorals (order Alcyonacea) to heat stress.In parallel, we evaluated the net changes in DOM (i.e., the concentrations of dissolved organic carbon [DOC] and dissolved organic nitrogen [DON]) uptake and release rates under normal and heat-stress conditions, to assess the resilience of DOM-supplied corals during heat stress.We compared these measurements to those of two species of hexacorals (order Scleractinia) maintained under the same experimental conditions.We used DOM derived from lysed zooplankton cells, which are likely highly labile.This is to mimic the capacity of corals to use DOM when it is freshly produced by plankton, or by "sloppy feeding" and extracellular enzymatic activities during coral grazing on plankton (Wijgerde et al. 2011), which can be very high in reef waters (Nichols et al. 2023).Our study addresses several questions.First, is the severity of bleaching and its effects on physiology and nutrition comparable in octocorals and hexacorals?Second, what are the fluxes of DOM under control conditions (26 C), and are they speciesspecific and concentration-dependent? Third, how do the fluxes of DOM change during heat stress (30 C)?We hypothesized that corals release DOM under normal growth conditions, whereas they take up DOM during heat stress.Furthermore, since octocorals have been shown to be more heterotrophic and more resistant to environmental stress than hexacorals, they should have milder bleaching and higher DOM uptake rates than hexacorals.A better knowledge of the dependence of different coral species on DOM is key to understanding the shift from hexacorals to octocorals in disturbed environments.

Experimental set-up
Two octocoral species (order Alcyonacea) (Sarcophyton glaucum, Quoy & Gaimard, 1833 and Lobophytum sp., Marenzeller, 1886) and two hexacoral species (order Scleractinia) (Stylophora pistillata, Esper 1797 and Turbinaria reniformis, Bernard, 1896) were selected for this experiment.For clarity, in the following sections as well as in the figures, S. glaucum and Lobophytum sp., will be referred to as Sarcophyton and Lobophytum, respectively, while S. pistillata and T. reniformis will be referred to as Stylophora and Turbinaria, respectively.The coral colonies originated from the Gulf of Aqaba, Red Sea, and were maintained in the running-seawater aquaria facilities of the laboratory.Six colonies of each species were used to produce 36 nubbins.All nubbins were placed in 6 Â 20-L independent aquaria (6 nubbins species À1 aquarium À1 from different colonies) maintained under the following conditions: a continuous supply of natural Mediterranean seawater pumped from 50 m depth at a flow rate of 20 L h À1 , with temperature kept constant at 26 AE 0.2 C using submersible resistance heaters (Visi-Therm VTX 300 W; Aquarium Systems) and temperature sensors (Ponsel).HQI lamps (Tiger 230 V/50 Hz 250 W; Faeber) provided photosynthetically active radiation (PAR) of 200 μmol photons À1 m À2 s À1 (12 : 12 light : dark) and submersible pumps ensured water mixing in each aquarium.Sarcophyton and Lobophytum nubbins were attached to a plastic net with nylon thread and placed on the bottom of the aquaria.The flattened shape of Turbinaria allowed the nubbins to be placed directly on the bottom, while Stylophora nubbins were glued to a plastic net using a two-component epoxy glue (HoldFast, Aquarium Systems) and placed on the bottom of the aquaria.Nubbins were fed Artemia salina nauplii twice a week, and tanks were cleaned weekly to avoid algal proliferation.After 8 weeks of acclimation, half of the tanks were identified as "control" with a stable temperature of 26 C, and the temperature of the remaining 3 tanks was increased by 0.4 C per day to 30 C over 10 d.Once the chosen temperature was reached, the effect of heat stress was monitored by measuring the effective photosynthetic efficiency (F v 0 /F m 0 ) of the corals using a pulse amplitude modulation (PAM) fluorometer (see later section).The experiment was stopped after 14 d at 30 C, when a significant decrease in this ratio was observed for all species (Table 1), suggesting that the photosynthetic capacity of the coral was impacted by heat stress.A decrease in F v 0 /F m 0 is also considered a precursor of bleaching (Jones et al. 1999).

Photosynthesis and respiration measurements
Rates of net photosynthesis (P n ) were assessed at 200 AE 10 μmol photons À1 m À2 s À1 on 6 nubbins per species and temperature.Rates of respiration (R c ) were measured in the dark, immediately after corals were maintained in the light for photosynthesis measurements.Nubbins were individually placed in 50-mL plexiglass chambers supplied with 0.45-μm filtered seawater and hermetically sealed.Stirring bars were used to ensure proper water homogenization.The chambers were equipped with a Unisense optode (oxygen-sensitive mini sensor) connected to the Oxy-4 software (Chanel fiber-optic oxygen-meter; PreSens).The optodes were calibrated with nitrogen-saturated and air-saturated seawater (for the 0% and 100% oxygen saturation levels, respectively) before each measurement.P n and R c were then determined by regressing oxygen values against time, and gross photosynthesis (P g ) was estimated by adding the absolute value of R c to P n .The daily percent contribution of photosynthesis to the holobiont respiratory requirements (referred to as P : R in the following sections) was calculated as (P/R) Â 100, where P represents the total daily grams of carbon per gram of ash-free dry weight (AFDW; see below) acquired through 12 h of gross photosynthesis (P g Â 12), and R represents the daily grams of carbon respired per gram of AFDW (R c Â 24).We assumed a moleto-mole ratio of CO 2 consumed (or produced) to O 2 produced (or consumed) during photosynthesis (or respiration).
For the control temperature, P n and R c were also assessed before and after the addition of DOC and DON to the chambers (reaching DOC 450 , DON 80 , see later section) for 6 nubbins per species, to evaluate the impact of high DOM concentration on photosynthesis and respiration.

Chlorophyll fluorescence measurements
The measurements were performed on the same nubbins used for photosynthesis assessments.A PAM fluorometer (Dual-PAM; Walz) was used to measure the minimal (F 0 0 ) and maximal (F m 0 ) chlorophyll a (Chl a) fluorescence of each nubbin in the light (at the culture irradiance) by applying a saturation pulse of light ($ 4000 μmol photons À1 m À2 s À1 , 0.8 s).
The effective photosynthetic efficiency of the photosystem II (PSII) was calculated as F v 0 /F m 0 = (F m 0 À F 0 0 )/F m 0 , according to Jones et al. (1999).This parameter provides a reliable indication of the PSII activity in dinoflagellates, and its decrease is related to PSII damage and/or a loss of efficiency of photochemical energy conversion, which can be considered a bleaching precursor signal (Jones et al. 1999).In addition, after 10 min of dark adaptation, the minimal (F 0 ) and maximal (F m ) Chl a fluorescence of each nubbin were also assessed by applying a saturation pulse of light.Thus, the maximum relative electron transport rate (rETR max ) was calculated by ([F m À F 0 ]/F m ) Â PAR Â 0.5, where the factor 0.5 assumes that half of the photons are absorbed by PSII (Genty et al. 1989).

DOC/DON fluxes experiments
For each species, incubations were carried out to evaluate the fluxes of DOC/DON at two temperatures (26 C and 30 C) and three DOM concentrations (ambient seawater and two enrichments with labile DOM) in a cross-factorial design.The DOM concentrations tested were the following: ambient seawater DOM concentration: 80 μmol C L À1 + 10 μmol N L À1 (DOC 80 , DON 10 ); a medium DOM enrichment resulting in a total concentration of 250 μmol C L À1 + 45 μmol N L À1 (DOC 250 , DON 45 ); and a high DOM enrichment resulting in a total concentration of 450 μmol C L À1 + 80 μmol N L À1 (DOC 450 , DON 80 ).Each combination of temperature and DOM concentration was performed under the culture irradiance (200 μmol photons À1 m À2 s À1 ) and in the dark (Fig. 1).
To produce the DOM used for the enrichment conditions, frozen microplankton (Ocean Nutrition™) was diluted in MilliQ water, homogenized at 20,500 rpm for 5 min using a high-performance dispersing instrument (T25 Ultra Turrax ® ; Janke & Kunkel), and then sonicated for 10 min.The homogenate was then centrifuged at 10,000 Â g for 1 h and the Table 1.Effective quantum yield (F v 0 /F m 0 ) values according to temperature (N = 6 for each condition) (* and *** mean significant differences with p < 0.05 and p < 0.001, respectively).Feeding of the nubbins was stopped 48 h before the incubations were performed (to avoid adding DOM or POM in the incubation medium through particulate feeding).The day before incubation, the tanks were also cleaned and a 50-μm mesh was added to the water inlet to limit the input of POM.Therefore, under these conditions, 95 AE 2% of the total organic matter was dissolved.We hereafter considered that all organic matter measured was DOM.At the beginning of the incubation, nubbins were individually transferred into 300 mL beakers placed beforehand in the aquaria.This prevents the corals from coming into contact with air, which could have resulted in abnormal mucus release.The beakers were then removed from the aquaria and incubated for 8 h on a stirring plate with magnetic stirrers (200 rpm) at the corresponding temperature, DOM, and light conditions applicable to the above-mentioned factorial experimental design.For each set of temperature, DOM concentration, and light level, three control incubations without nubbins were also done.
At the beginning (t 0 ) and end of the incubation (t 8 ), seawater samples (15 mL) were collected in duplicate for dissolved inorganic nitrogen (DIN) measurements using an auto-analyzer according to Treguer and Le Corre (1975) (see full description in Supporting Information Material) to calculate DON from total dissolved nitrogen (TDN; DON = TDN À DIN).Thirty-milliliter seawater samples were also collected in duplicate from each beaker using a sterile syringe, then transferred to a precombusted (4 h at 450 C) Pyrex ® tube, and stored at À20 C. The concentrations of DOC and TDN in the seawater samples were determined using a TOC-L analyzer (Shimadzu) (see full description in Supporting Information Material).The net DOC/DON fluxes were calculated using the difference between the DOC/DON concentration measured at t 8 and the concentration measured at t 0 , after correction for seawater controls.The net DOC/DON fluxes were then normalized to the incubation time and AFDW of each nubbin, so all fluxes are expressed in μmol C or N g À1 AFDW h À1 .Negative fluxes indicated net uptake rates of DOC/DON, whereas positive fluxes indicated net release rates by the coral.At the end of incubations, nubbins were frozen and stored at À20 C.

Contribution of DOC heterotrophy to coral respiration
The contribution of DOC heterotrophy to the holobiont respiration (CHAR DOC , %) was calculated as CHAR DOC = (DOC f / R c ) Â 100, DOC f being the daily carbon acquisition by DOC heterotrophy and was calculated as the sum of the light and dark DOC fluxes (assuming a 12 h : 12 h light : dark cycle).

Sample analysis and data normalization
Nubbins of Sarcophyton and Lobophytum were directly freezedried to determine their dry weight (DW) (Pupier et al. 2018).For the hexacorals, tissue was first removed from the skeleton using an airbrush and homogenized with a Potter tissue grinder (Ezzat et al. 2016) before being freeze-dried.The samples were then processed as described in Pupier et al. (2018).Briefly, tissue samples were ground to a powder in a mortar, a fraction was taken for AFDW determination, while the remaining powder was homogenized with 10 mL of MilliQ water.Two subsamples of 500 μL were collected for the measurements of protein content and symbiont density (see next section).Finally, 5 mL of the homogenates were taken for the determination of the total chlorophyll concentration (next section).All data were normalized to the nubbin's AFDW, according to Pupier et al. (2018).

Tissue parameters measurements
The genus of Symbiodiniaceae associated to each coral species was determined, following the protocol of Santos et al. (2002).Symbiont density was measured through microscope counting via eight replicate hemocytometer counts (Neubauer hemocytometer; Marienfeld).For chlorophyll quantification, the collected subsamples were centrifuged at 8000 Â g and 4 C for 10 min.The supernatant was removed, and the pellet was washed twice with MilliQ water.Then, 5 mL of acetone (> 99%; Sigma-Aldrich) amended with magnesium chloride (Sigma-Aldrich) were added to the symbiont pellet, to extract chlorophyll for 24 h in the dark at 4 C.After centrifugation at 11,000 Â g for 15 min at 4 C, absorbances at 616 nm and 663 nm were measured using a Xenius spectrophotometer (SAFAS ® ) and the total chlorophyll concentration was determined according to Jeffrey and Humphrey (1975).Proteins were extracted in sodium hydroxide solution (1 mol L À1 ) for 30 min at 90 C. Protein content was then quantified using a BC assay kit (Interchim) (Smith et al. 1985) and protein standards were prepared using bovine serum albumin (Interchim).The absorbance was measured at 562 nm by a Xenius spectrophotometer (SAFAS ® ).
The total organic carbon (C) and nitrogen (N) contents of the tissue were analyzed on acidified samples in tin capsules containing 600 μg of freeze-dried tissue using an Integra II mass spectrometer (Sercon ® ).To determine the phosphorus (P) content, sodium chloride (4 mL, 2 N) and potassium persulfate (68 μL, 50 g L À1 ) were added to the freeze-dried tissue in Pyrex ® tubes before the liquid was autoclaved.Solutions of chlorohydric acid (33 μL, 0.6 N) and sodium bicarbonate (268 μL, 0.035 N) were then added.A solution of AMP was prepared with sulfuric acid (1.06 mL) and ammonium molybdate (175.5 mg) dissolved in 10 mL of MilliQ water.Then 150 μL of the homogenate to be measured were placed on a microplate along with 30 μL of the AMP solution, and malachite green (30 μL, 0.35 g L À1 ) was added after 10 min.Phosphorus standards were prepared with potassium phosphate, and absorbance was measured at 630 nm using a Xenius spectrophotometer (SAFAS ® ).

Statistical analysis
Analyses were performed using R software (R Foundation for Statistical Computing, version 4.0.2).All data were expressed as mean AE standard deviation.Before performing the analyses, outliers were identified using Grubb's test and were excluded when p values were significant (p < 0.05).The assumptions of normality and homoscedasticity of variances were assessed using Shapiro's and Bartlett's tests along with graphical analyses of the residuals.When they were satisfied, a two-way analysis of variance was performed to test the effect of species and temperature on F v 0 /F m 0 , rETR max , chlorophyll content, CNP content, net/gross photosynthesis, and respiration rates.When there were significant differences between treatments, a posteriori test was performed (Tukey's HSD test).
For the effect of DOM enrichment on net/gross photosynthesis and respiration rates, a Student t-test pair was performed.
Otherwise, when the assumptions were not satisfied, a Kruskal-Wallis test was used, looking separately at the effect of temperature and species on protein content, Symbiodiniaceae density, gross photosynthesis rates, P : R and CHAR DOC .For the DOC/DON fluxes, the separate effect of DOC/DON concentration and irradiance condition was also considered.When the effect was significant, a posteriori test was performed (Dunn's test).For all statistical tests, differences were considered significant when p < 0.05.For the DON measurements, there was no significant effect of irradiance (light vs. dark; see Supporting Information Table S2) on DON fluxes, therefore, the results were pooled.

Physiological and tissue differences between species at 26 C
The results of the statistical tests are reported in Supporting Information Table S3 for physiological parameters and in Supporting Information Table S4 for photosynthesis and chlorophyll fluorescence measurements.
All coral species harbored Cladocopium sp.(> 99.31% identity similarity), except the hexacoral Stylophora, which was associated with Symbiodinium sp.(> 98.94% identity similarity).Protein content at 26 C was not significantly different between species ( p > 0.05; Fig. 2a; white boxplots in figures).However, chlorophyll content per AFDW was ca.threefold lower in octocorals than in hexacorals (p < 0.05; Fig. 2b).This was due to an overall lower symbiont density in octocorals, although the difference was significant only when compared to the hexacoral Stylophora ( p < 0.05; Fig. 2c).As a result of the lower chlorophyll content, the net (P n ) and gross (P g ) photosynthetic rates of octocorals were three times lower than those of Stylophora (p < 0.001 and p < 0.05, respectively; Fig. 3a,b), in agreement with lower rETR max (p < 0.01; Fig. 3d).Despite the low P n and P g rates, the P : R ratio (the contribution of photosynthesis to the respiratory requirements, shown in Supporting Information Fig. S1) was above 100% in octocorals, as rates of respiration were also low compared to hexacorals (p < 0.05; Fig. 3c; note the negative sign on the y-axis).The hexacoral Turbinaria was the only species that did not achieve a P : R > 100%.Moreover, octocoral Sarcophyton was the only species whose photosynthetic rates were increased by DOM enrichment (average +44%; p < 0.01; Supporting Information Fig. S2), while its respiration rates were not significantly affected.In the octocoral Lobophytum, photosynthetic rates remained constant, but respiration rates increased with DOM enrichment (average +21%; p < 0.05).For hexacorals, DOM enrichment also had no effect on photosynthetic rates, although it slightly reduced respiration rates (average À20% for Stylophora and À5% for Turbinaria, p < 0.01).
In terms of elemental tissue composition, octocoral species had higher organic C content than hexacorals (p < 0.01; Supporting Information Fig. S3a), but overall, there was no significant difference between species in N content (Supporting Information Fig. S3b), or C : N ratios (Supporting Information Fig. S3d).Turbinaria had the lowest P content (p < 0.001; Supporting Information Fig. S3c).

Effect of heat stress on coral physiological and tissue parameters
The results of the statistical tests are reported in Supporting Information Table S3 for physiological parameters and in Supporting Information Table S4 for photosynthesis and chlorophyll fluorescence measurements.
The effect of heat stress was species-specific (black boxplots in the figures).No effect of temperature on the protein content, chlorophyll content, symbiont density, and photosynthesis was observed in octocorals ( p > 0.05; Figs.2a-c, 3a,b), except for P n of Lobophytum, which decreased by almost 40% under heat stress ( p < 0.01), in parallel to a 20% decrease in P : R ( p < 0.01; Supporting Information Fig. S1).However, octocorals presented a significant increase in rETR max under heat stress (35% and 63%, respectively; p < 0.01; Fig. 3d).For the hexacoral Stylophora, heat stress significantly increased chlorophyll and protein concentration (p < 0.001 and p < 0.01, respectively; Fig. 2a,b), which resulted in no significant change on P n and P g (p > 0.05; Fig. 3a,b), despite a significant decrease of rETR max (15%; p < 0.01; Fig. 3d).Under heat stress, the hexacoral Turbinaria lost 25% of its symbionts and 30% of its chlorophyll content (p < 0.05).This led to a 40% and 20% decrease in P n and P : R (p < 0.01 for both), respectively, due to increased respiration rates (p < 0.05; Fig. 3c).The rETR max also significantly decreased (by 65%; p < 0.01; Fig. 3d) and was significantly lower than that of the other species (p < 0.05).
In terms of elemental tissue composition, no significant changes in C, N, and P content were noted for octocorals (Supporting Information Fig. S3a-c).On the contrary, there was a significant increase in tissue N for Stylophora and an increase in tissue P for Turbinaria (p < 0.01 and p < 0.001, respectively; Supporting Information Fig. S3b,c).

Changes in DOC fluxes as a function of environmental parameters
The results of the statistical tests are reported in Supporting Information Table S5 for DOC fluxes and in Supporting Information Table S6 for CHAR DOC measurements.
At ambient seawater DOC concentration (DOC 80 ), net DOC fluxes were close to zero for all corals, temperatures, and light levels (balanced uptake and release rates; Fig. 4; Supporting Information Table S1).At higher DOC concentrations (DOC 250 and DOC 450 ), there was a net uptake of DOC by all coral species (p < 0.05).Uptake rates varied depending on the species, temperature, or irradiance, but overall, Lobophytum had the lowest uptake rates of all four species.Heat stress had no effect on DOC uptake rates of Sarcophyton, but reduced those of Lobophytum at the highest DOC concentration (DOC 450 ; p < 0.01).In parallel, heat stress significantly increased DOC uptake rates of Stylophora at DOC 250 and DOC 450 (p < 0.05), while the opposite was observed for Turbinaria (p < 0.05).Finally, light levels had no effect on DOC uptake in octocoral species, in contrast to hexacorals.Indeed, DOC uptake rates of Stylophora were significantly higher in the dark, for both DOC 250 and DOC 450 at 26 C, as well as for DOC 250 at 30 C (p < 0.05).Similarly, DOC uptake rates of Turbinaria were significantly higher in the dark for DOC 450 at 26 C, and for DOC 250 at 30 C (p < 0.05).
At 26 C, the contribution of heterotrophic DOM-feeding to the total coral respiration (CHAR DOC ; Table 2) followed the same trends as the DOC fluxes.It significantly increased with DOC supply in all four species (p < 0.05), and even reached a value of 160% at DOC 250 for Sarcophyton.Therefore, all species reached a CHAR DOC > 100% at DOC 450 , except for Stylophora.Heat stress had no effect on CHAR DOC in Sarcophyton, but reduced CHAR DOC in Lobophytum at DOC 450 (p < 0.05).On the contrary, there was a significant increase in CHAR DOC for Stylophora under heat stress, for both DOC 250 and DOC 450 (p < 0.05).Indeed, at DOC 450 , its CHAR DOC doubled from 80% at 26 C to 160% at 30 C.

Changes in DON fluxes as a function of environmental parameters
The results of the statistical tests for DON fluxes are reported in the Supporting Information Table S7.At ambient seawater DON concentration (DON 10 ), net DON fluxes were close to zero (uptake and release balanced; Fig. 5), but DON enrichment led to net DON uptake in all species investigated at 26 C and 30 C (p < 0.01 and p < 0.05, respectively).Irradiance (light vs. dark) had no significant effect on nitrogen fluxes for both octocorals and hexacorals (Supporting Information Table S2).

Discussion
Octocoral assemblages create three-dimensional structures of reefs (S anchez 2017), and are one of the major alternative groups of the benthic community of reefs that have diverged from hexacoral dominance, particularly in the Western and Central Indo-Pacific regions (Reverter et al. 2022).This suggests that octocorals may be more resilient, or better able to adapt to changes in biotic and abiotic environmental conditions in future reefs.Such resilience may be explained by a higher dependence on heterotrophy, in particular on DOM as a food source.However, only a few studies exist on DOM fluxes in octocorals.Therefore, the main objective of this study was to compare the dependence of octocorals on DOM in their mixotrophic diet, under different environmental ).AFDW, ash-free dry weight.Significant differences between temperatures or irradiance conditions for each species are displayed with asterisks (p < 0.05).Significant differences in DOC fluxes between the lowest and highest DOC concentration (in the dark, at 200 μmol photons À1 m À2 s À1 , both conditions) are displayed with an arrow (black continuous, gray dotted, black dotted, respectively).
conditions, and compare it with that of hexacorals.The main results obtained show that the feeding strategy of octocorals on DOM is not different from that of hexacorals.For all coral species investigated, net DOC/DON uptake rates were positively correlated to DOM concentrations in seawater.In addition, we observed that species that were able to increase (the hexacoral Stylophora) or maintain (the octocoral Sarcophyton) their feeding rates on DOM at high temperature were also the most resistant species to heat stress.In contrast, species that decreased their DOM uptake rates (the hexacoral Turbinaria and, to a lesser extent, the octocoral Lobophytum) showed higher susceptibility to heat stress.

DOC/DON fluxes under control conditions (26 C)
Ambient DOM concentrations are relatively low in most reefs (de Goeij and van Duyl 2007; Haas et al. 2011;Tanaka et al. 2011), and corals tend to be net producers of DOM (Ferrier-Pagès et al. 1998;Naumann et al. 2010;Haas et al. 2011;Levas et al. 2015), with the exception of some branching species, such as S. pistillata and Pocillopora sp. that were found to sometimes be net consumers (Naumann et al. 2010;Tremblay et al. 2012b).Our four tested species, including octocorals, were indeed net producers under ambient DOM concentrations, in the light, and at control seawater temperature.However, all coral species changed from net producers to net consumers of DOM after the ambient water was enriched with labile DOM.Indeed, our four species tested were increasingly consuming DOC/DON at higher DOM  concentrations in seawater.Concentrations as high as those used in our middle enrichment condition (DOC 250 , DON 45 ) can occur in mangroves or coastal waters exposed to groundwater or river discharge (Rezende et al. 2007;Coni et al. 2017;Webb et al. 2019).Healthy corals, such as those of the Great Amazon Reefs, can thrive in these environments (Francini-Filho et al. 2018;Camp et al. 2019), and have even been shown to be particularly resistant to environmental stress thanks to a high degree of heterotrophy (Camp et al. 2019) in their mixotrophic diet.However, the coral diversity in these reefs is lower than in oligotrophic reefs, and further studies should aim to investigate the interactions between corals and macroalgae in such DOM-concentrated environments.In addition, the DOM used in this study was likely highly labile and directly bioavailable by corals, because it was derived from lysed zooplankton cells.This is not the case with mangroves or river discharge, where the majority of DOM is recalcitrant and is not bioavailable to corals.Nevertheless, labile DOM can be locally available in high quantities while corals graze on zooplankton through sloppy feeding (Wijgerde et al. 2011), or through extracellular enzymatic activities.These enzymes can be found at high concentrations in reefs since they are continuously released by corals and microbes, and contribute largely to the degradation of particulate and dissolved matter (Nichols et al. 2023).
Our study, however, highlights species-specific differences in nutritional strategies.The most efficient user of DOM was the octocoral Sarcophyton, which was the only species that entirely met its metabolic needs from DOM-heterotrophy at DOC 250 , with a CHAR DOC well over 100%.Together with the carbon gained through photosynthesis (P : R), this species met more than four times its respiratory requirements.Moreover, it was the only species whose photosynthesis was increased at high DOM enrichment.This finding is interesting, as Sarcophyton is often a dominant species in degraded and DOM-concentrated environments (Rodriguez et al. 2020;Lalas et al. 2021), while it does not grow when supplied with particulate food in the form of rotifers (Costa et al. 2016).Therefore, this species seems to be particularly well adapted to feed on dissolved material.The other octocoral Lobophytum and the hexacoral Turbinaria also reached 100% of their respiratory needs when fed DOC 250 , in addition to the carbon gained through photosynthesis.Feeding on DOM can be particularly important for Turbinaria, as it was the only species unable to compensate for its metabolic demand through photosynthesis.This is in agreement with the observation that Turbinaria tends to thrive in turbid and organically rich environments (Anthony 2006;Ezzat et al. 2016).Finally, Stylophora tended to have high DOM uptake rates, but its CHAR DOC remained low (< 100%) due to an equally high respiration.It was, therefore, the only species that, even at the highest DOM concentration, did not feed enough on DOM to bring its CHAR DOC to 100% and meet its daily metabolic demand only from DOM-heterotrophy.This species may be better adapted to capture particulate prey, such as zooplankton.
Interestingly, under ambient DOM concentration, the hexacorals shifted toward the net uptake of DOC in the dark.The apparent higher DOC consumption may just be the result of lower release rates of DOC by corals during the night, shifting the balance between uptake and release toward net uptake.Indeed, in algae, but also corals, the release of photosynthetically fixed carbon in the form of DOC has been shown to be directly related to primary production, with a positive relationship between DOC release and light availability (e.g., Kurihara et al. 2018;Mueller et al. 2022).During the day, photosynthesis can meet more than 100% of the respiratory needs of hexacorals, which therefore release the excess carbon as DOC and mucus (Davies 1984).Unlike hexacorals, octocorals did not increase their DOC uptake rates in the dark, likely due to overall lower rates of gross photosynthesis compared to hexacorals (Fabricius and Klumpp 1995;Pupier et al. 2019).In contrast to DOC, there were no effects of light/dark conditions on net uptake rates of DON, in both hexacorals and octocorals studied here.Since nitrogen is a limiting nutrient for coral growth, it is highly retained within the symbiotic association through internal recycling (Wang and Douglas 1999).Several studies have shown that DON is rapidly taken up when present in seawater, mainly in the form of dissolved free amino acids (DFAAs) or urea (Hoegh-Guldberg and Williamson 1999;Grover et al. 2006Grover et al. , 2008;;Gori et al. 2014;Pupier et al. 2021) with the exception of dissolved combined amino acids, which can be released due to their larger molecular size (Tanaka et al. 2009).However, DFAAs are highly labile and might not represent the full spectrum of DOM molecules found in natural environments.It is also important to note that the DOM used in this study was generated from natural plankton assemblages and may have been more labile than the natural DOM pool in reef waters.
Although the net fluxes of DOM show a net release in the light and at ambient DOM concentrations, the comparison between light and dark fluxes estimates that corals in the light and at ambient DOM concentration took up between 8.8 and 27.8 μmol C g À1 AFDW h À1 and 0.9-2.0μmol N g À1 AFDW h À1 .These calculations are in agreement with previous measurements performed with 15 N-labeled DFAAs, leading to uptake rates of 0.3 to 1 μmol N g À1 AFDW h À1 at similar DON concentrations in seawater.Compared to the rates of photosynthesis measured in this study, or to the rates of inorganic nitrogen uptake measured in previous studies (Grover et al. 2002(Grover et al. , 2003;;Pupier et al. 2021), these low rates of DOM uptake can supply a significant amount of carbon and nitrogen to the corals.However, further studies under ambient DOM conditions are needed to fully understand the contribution of DOM to the carbon and nitrogen budgets of corals.

Effect of heat stress on coral physiology and DOM uptake rates
Under heat stress, most symbiotic corals bleach (i.e., decrease in symbionts and/or chlorophyll pigments), and experience a general decrease in their autotrophic acquisition of nutrients (Tremblay et al. 2012b;Rädecker et al. 2021).It has, therefore, been suggested that corals that can increase their heterotrophic nutrient uptake will be among the winners under future climate change conditions (Grottoli et al. 2006;Towle et al. 2015;Aichelman et al. 2016).It is, therefore, of prime importance to assess the full heterotrophic capacities of different coral species under stress conditions.However, and although some studies have examined the reliance of corals on particulate food or on the total organic matter present in seawater (Ferrier-Pagès et al. 1998;Naumann et al. 2010;Haas et al. 2011;Tremblay et al. 2012b;Levas et al. 2015), to our knowledge only three have investigated the feeding of corals on DOM under heat stress.These studies have found that some hexacoral species (Porites divaricata, Porites astreoides, Orbicella faveolata) can increase their reliance on DOC when bleached, whereas other species (such as T. reniformis, Acropora millepora, and Acropora muricata) cannot (Levas et al. 2015(Levas et al. , 2016;;Courtial et al. 2017).Our results are in line with these few previous observations since the two octocoral species did not change their uptake rates of DOM with heat stress at most DOM concentrations, while changes did occur in the hexacoral species, although with different trends.A significant decrease in DOM uptake rates was indeed observed in Turbinaria, while a significant increase occurred in Stylophora.
Our results further highlight similar trends between the DOM fluxes and changes in coral physiology.Indeed, species that were able to increase (Stylophora) or maintain (Sarcophyton) their feeding rates on DOM were also the most resistant species to heat stress (see below).On the contrary, species that decreased their DOM uptake rates (mainly Turbinaria and Lobophytum at the highest DOM concentration) showed higher susceptibility to heat stress.However, whether decreases in DOM uptake rates explain the overall decline in coral physiology or the opposite remains to be investigated.Stylophora is the only thermophilic species, as at high temperature, it significantly increased its chlorophyll and protein concentrations (and therefore cellular N content) as well as uptake rates of DOC and DON while maintaining high rates of net and gross photosynthesis compared to the other species.Stylophora is an emblematic species of the Gulf of Aqaba (Red Sea) and is known to have an exceptionally high bleaching threshold, well above the maximum temperature recorded for the region (Bellworthy and Fine 2017;Evensen et al. 2021).Its high level of heterotrophy on DOM, combined with its thermophilic behavior, allowed this species to have a higher total acquired fixed carbon relative to the total coral respiration (CTAR) at 30 C than at 26 C, and to likely be a winner under the future climate conditions.For Sarcophyton, CTAR always remained above 100%, likely due to a high degree of heterotrophy on DOM when in a DOM-enriched environment, well above its autotrophic capacity.The absence of a negative response to heat stress might, however, be due to the higher bleaching threshold of this genus.Bleaching has been regularly reported for this genus, but at temperatures above 30 C (Marshall and Baird 2000;Chavanich et al. 2009).Although there were no major changes in its physiology, the C : N and C : P ratios of the tissue of Sarcophyton significantly increased under heat stress, suggesting a nitrogen and phosphorus limitation of tissue growth.Finally, the other two coral species, mainly Turbinaria and to a lesser extent Lobophytum, that decreased their feeding rates on DOC/DON under heat stress were also the species most impacted by heat stress, with a decrease in photosynthetic capacity and/or significant bleaching, and a CTAR below 100% at ambient DOM concentration (Turbinaria).The increase in cellular P content in Turbinaria may suggest a high level of oxidative stress, and the mobilization of polyphosphates as non-protein antioxidant molecules (McCain and Bertrand 2022).Overall, these physiological and feeding results suggest various degrees of heterotrophy on DOM in the coral species investigated, coupled to various heat-stress resistances.

Conclusion
DOM, which includes organic molecules such as polysaccharides, proteins, and lipids, is a key component of biogeochemistry and overall reef functioning because it is a major source of organic carbon and nitrogen to many reef organisms (Haas et al. 2011).DOM is often converted into organic particles (POM) via the microbial loop (Azam et al. 1983) and the sponge loop (de Goeij et al. 2013) before being transferred to the higher trophic levels.Sponges, for example, can remove up to 340 mmol DOC m À2 d À1 at ambient DOC concentrations in seawater, which is in the same order of magnitude as gross primary production rates of an entire reef ecosystem (de Goeij and van Duyl 2007;de Goeij et al. 2008).Although corals are considered DOM producers rather than consumers, this study's estimates derived from light and dark measurements suggest that average coral biomass in reefs of 5 mg AFDW cm À2 (Thornhill et al. 2011), can remove 1-3 mmol DOC m À2 d À1 .Furthermore, the results of this study show that all coral species investigated can feed significantly on DOM when it is present in sufficient amounts in seawater.This is the case for reefs near river plumes (as in the Amazon system), mangroves, or eutrophied coastal environments, all of which are usually enriched in DOM (Tedetti et al. 2011;Moura et al. 2016) and host corals that are particularly resistant to environmental stress.Although we have shown that all coral species studied are able to take up DOM, further studies are needed to specifically investigate the ability of species living in DOM-enriched environments to feed on DOM.The results obtained also reinforce that there are great differences between coral species in their capacity to use DOM.Therefore, generalizing across corals as a monolith may be unwise for researchers and stakeholders with an interest in preserving and restoring reefs.Finally, concentrations of DOM in reefs are predicted to increase in the future as turf and macroalgae abundance will increase due to climate change disturbance (de Goeij et al. 2017).In addition, not only the quantity but also the quality (i.e., bioavailability) of DOM will change depending on the main producers, which will affect the overall biodiversity and function of the reef.Therefore, to better predict future reef biodiversity, further studies should investigate the ability of corals to feed on different DOM sources, compared to sponges and other reef organisms.
filtered through a 0.20-μm filter before being stored at À20 C. The resulting labile DOM had a concentration of 100 mmol C L À1 and 20 mmol N L À1 .
Fig. 1.Overview of the experimental design for measurements of dissolved organic matter (DOM) fluxes.Corals were either maintained at 26 C or 30 C in replicated aquaria.The seawater used for the incubations was either not enriched (ambient DOM concentration: 80 μmol C L À1 + 10 μmol N L À1 ), moderately enriched with DOM (250 μmol C L À1 + 45 μmol N L À1 ) or highly enriched (450 μmol C L À1 + 80 μmol N L À1 ).Sarcophyton glaucum and Lobophytum sp. are octocorals, while Stylophora pistillata and Turbinaria reniformis are hexacorals.All incubations were performed under an irradiance of 200 μmol photons À1 m À2 s À1 and in the dark.

Fig. 4 .
Fig. 4. Dissolved organic carbon (DOC) fluxes (μmol C g À1 AFDW h À1 ) under different temperatures (26 C and 30 C), irradiance and DOC concentrations.White striped bars represent corals under an irradiance of 200 μmol photons À1 m À2 s À1 , while black bars represent corals in the dark.From left to right: ambient DOC concentration: 80 μmol C L À1 ; moderate DOC enrichment (250 μmol C L À1 ) and high DOC enrichment (450 μmol C L À1). AFDW, ash-free dry weight.Significant differences between temperatures or irradiance conditions for each species are displayed with asterisks (p < 0.05).Significant differences in DOC fluxes between the lowest and highest DOC concentration (in the dark, at 200 μmol photons À1 m À2 s À1 , both conditions) are displayed with an arrow (black continuous, gray dotted, black dotted, respectively).

Table 2 .
Contribution of heterotrophy on dissolved organic carbon (DOC) relative to the holobiont respiration (CHAR DOC , %) regarding species, temperature, and DOC concentration.