Sample collection: Fieldwork was conducted at Mote Marine Laboratory in Summerland Key, FL, USA, between July 27 and August 19, 2021. Replicate individuals of the four sponge species Aplysina cauliformis, Iotrochota birotulata, Niphates digitalis, and Xestospongia muta were collected from a shallow (7 m depth) patch reef (24°33’ N, 81°24’ W) offshore of Summerland Key (24°39’ N, 81°26’ W) via dive knives (with the spongocoel intact for the barrel/tube sponges by cutting the sponge from the substrate) using SCUBA, and were placed into plastic bags filled with seawater before being brought to the surface. After collection, all sponges were kept submerged in ambient seawater in large, sealed plastic bags in an insulated cooler during transit back to the lab. Upon return to Mote Marine Laboratory, sponges were attached with cable ties to plastic window screens, and held in large, flow-through raceways for 24 h for acclimatization. At no point during this process were sponges exposed to air.

Only healthy sponges (sponges confirmed to be actively pumping via use of fluorescein dye, and without any signs of necrosis) were processed for and used in experiments. For each experiment (and each treatment in the autotrophy experiment), four large individuals of each of the rope sponges I. birotulata and A. cauliformis collected from the reef were partitioned into seven subsamples for sampling at discrete time points. Both X. muta and N. digitalis were only sampled at four timepoints during each experiment because they have barrel/tube morphologies that cannot be successfully partitioned and subsampled, requiring the sacrifice of an entire individual for each timepoint. Therefore, sixteen small individuals of each of these two species were collected from the reef for each experiment. Experimental replicates were defined as follows: a replicate of X. muta consisted of one individual attached to a plastic window screen, which was placed in an 8 L, clear food-grade container; a replicate of N. digitalis consisted of four small individuals attached to a plastic window screen, which was placed into a 17.9 L, clear food-grade container; and replicates of A. cauliformis and I. birotulata consisted of all seven subsamples from the same individual attached to a plastic window screen, which was placed into a 17.9 L, clear food-grade container. In total, there were four replicates of each species collected at each time point in each experiment, with the exception the DOM experiment, during which multiple individuals, including all individuals of A. cauliformis, developed necroses in experimental tanks after 24h. The effected individuals were eliminated, limiting replication to only two timepoints (24h and 72h) during the chase for all species. Due to necroses of all individuals of A. cauliformis after 24h, the DOM experiment was repeated for this species only, using three healthy replicates.

Pulse-chase experiments: Three separate “pulse-chase” experiments were carried out to track the uptake and recycling of labeled compounds from three different resource pools. The first resource pool, DOM, consisted of ¹³C-labeled glucose (35 μM) and galactose (35 μM), and ¹³C and ¹⁵N-labeled cell-free amino acids (range from 39-790 nM) and urea (2.2 μM). The second resource pool, POM, consisted of ¹⁵N- and ¹³C-labeled Synechococcus spp. bacteria incubated for 36 h in seawater containing 1.18 μM NaH¹³CO₃, 3.67 μM ¹⁵NH₄Cl, and 0.117 M Na¹⁵NO₃, and Oceanicola batsensis incubated in 0.55 μM galactose, 1.18 μM NaH¹³CO₃, 3.67 μM ¹⁵NH₄Cl, and 0.117 μM Na¹⁵NO₃. Finally, the third resource pool of inorganic compounds consisted of Na¹⁵NO₃ (0.117 μM) and NaH¹³CO₃ (1.18 μM). Labeled bacterial cells were added to filtered seawater for the POM experiment at a final combined concentration of 4.65 x10⁵ cells ml^-1 for photosynthetic cells and 1.01 x 10⁶ for heterotrophic cells (measured at the Center for Aquatic Cytometry at Bigelow Laboratory for Ocean Sciences). Concentrations of labeled resources were chosen to approximate the concentrations of corresponding inorganic and organic sources of carbon and nitrogen found in Caribbean seawater as closely as possible, as well as to match concentrations used in previous studies. In all incubations, experimental water was 0.7 μm (GF/F) filtered prior to the addition of tracers or labeled bacteria. Sponges were “pulsed” for 3 h in water containing these isotopically enriched resources, and then sponges were held in unenriched, flowing seawater for a 72-h “chase” period. Experimental samples were collected during the pulse at t=0.5 h, and during the chase at final times of t=12 h, 48 h, and 75 h (minus the 48h sampling in the DOM experiment). Prior to the start of the pulse, one of the seven subsamples of A. cauliformis and I. birotulata, and four individuals each of X. muta and N. digitalis were sampled and processed to serve as t = 0 h (natural abundance) samples.

To start the pulse, each experimental replicate was added to a food grade container containing filtered seawater and labeled compounds or food bacteria, as well as a recirculating pump used to maintain flow. Replicates of X. mutawere held in 8 L, clear food grade containers holding 6 L of filtered seawater, while replicates of I. birotulata, A. cauliformis, and N. digitalis were held in 17.9 L containers holding 15 L of filtered seawater. These containers were all held in a large 4.6 m-diameter outdoor tank with a shallow seawater bath to control temperature in experimental tanks holding sponges. At the end of the pulse, the water level in the outdoor tank was raised to flood experimental tanks with unlabeled seawater (supplied from the canal adjacent to Mote Marine Lab, drawn from 10 m, and filtered through sand to 40 μm) throughout the chase. Temperature measurements were taken at regular intervals (30 min), and an additional 25°C supply of well water, with salinity equal to that of seawater, was supplemented to maintain a tank temperature between 29-32°C. During the day for all experiments, irradiance measurements were taken every 30 min to monitor for deviation outside of the irradiance that sponges experience at depth, which was 800 to 1000 μmol photons m^-2 s^-1 measured at depth during cloudless daytime conditions at the collection site. One layer of standard shade cloth was sufficient to keep irradiance within the normal limits of what the sponges experience at depth, and adjustments during the experiment were only made when irradiance went below the 800-1000 μmol photons m^-2 s^-1 range under cloudy conditions. When irradiance decreased, we removed the layer of shade cloth to increase exposure of the sponges to ambient sunlight.

To assess the role of both autotrophic and heterotrophic symbionts in the assimilation of inorganic compounds, the autotrophy experiment included both light and dark treatments. For this reason, each large individual of I. birotulata and A. cauliformis was partitioned into 14 subsamples, and 32 small individuals each of N. digitalis and X. muta were collected for this experiment. Half of the subsamples or individuals of each species were randomly assigned to a dark treatment (covered in opaque plastic tarp with irradiance values from HOBO loggers that detected 0 Lux throughout the pulse), and the other half to a light treatment (exposure to ambient irradiance as in other experiments). Samples in the dark treatment were covered during the pulse but were exposed to ambient irradiance during the chase so that we could monitor nutrient retention and recycling under ambient (light-dark cycle) conditions.

Sample fixation: After sampling during pulse-chase experiments, sponges were rinsed with 0.7 μm-filtered, unenriched seawater and lightly blotted with a paper towel before weighing to the nearest 0.001 g to obtain a wet weight. Afterwards, ~2 mm cross sections were collected in duplicate using a razor blade and placed into 2 mL cryovials to be fixed for SEM and NanoSIMS analysis; the remaining bulk tissue was frozen at -20°C for isotope analysis. The 2 mm cross-sections of sponge tissue were fixed in 2.5% (w/v) glutaraldehyde + 1% (w/v) paraformaldehyde in 1.5X PHEM buffer (1.5X PHEM (90 mM PIPES, 37.5 mM HEPES, 15 mM EGTA, 3 mM MgSO4.7H₂O), and 9% (w/v) sucrose at pH 7.4) and stored at 4°C for 12h before being rinsed 3x with PHEM buffer containing no fixative. Finally, 0.1 mL of fixative solution (10% of the concentration in the original solution) was added back to each sample to prevent any fouling during long-term storage at 4°C. After being taken out of storage, cross-sections of sponge tissue were rinsed once more in fixative-free 1.5X PHEM buffer, digested in 5% hydrofluoric acid to break down siliceous spicules within the sponge tissue (note this was carried out for all samples to be consistent, even though A. cauliformis lacks spicules), and rinsed twice and stored in 1.5X PHEM buffer.