Experiment Conditions and Details of Resulting Data

Pelagic Sargassum incubations were conducted at the Bermuda Institute of Ocean Sciences (BIOS) during the summers of 2018 and 2019. Samples were collected from 4 km northeast of Bermuda to ensure samples reflected offshore conditions. Following established protocols (Hanson, 1977; Lapointe, 1986; Schofield et al., 1998; Smith et al., 1973), Sargassum patches were collected via dip nets and placed in a cooler for transport back to land. The Sargassum samples were transferred to outdoor aquaria in full sunlight with flowing seawater taken from the adjacent bay to maintain ambient surface seawater temperatures until ready for use in incubations. These samples remained in the aquaria for no more than one full day before use and were only selected if in visibly good health. 

For experiments, 1-2 g (wet weight) Sargassum samples (stipe, blades, and pneumatocysts) were clipped, rinsed to remove large macrofauna (Hanson, 1977; Lapointe, 1995), and placed into 150 mL borosilicate glass serum bottles. Serum bottles were then filled with 75 mL (in 2018) or 125 mL (in 2019) autoclaved seawater from the bay adjacent to BIOS, leaving sufficient headspace for subsampling. Note that this design eliminates the contribution of the pelagic bacterial community associated with the macroalgae. The components for specific treatments (including MPn, ammonium nitrate, sodium phosphate, and an antibiotic mixture containing 50 µg/mL ampicillin, 50 µg/mL kanamycin, 30 µg/mL nalidixic acid and 100 µg/mL streptomycin) were added from stock solutions, then the bottles were sealed with N-butyl rubber stoppers and crimp caps.

To determine how much MPn to add to treatments, we estimated the amount of MPn in the region (further details related to how MPn concentrations were determined can be found at the end of this Methods & Sampling section of this metadata page). Lomas et al., (2009) determined that 80% of total dissolved phosphorus (TDP) is in the form of dissolved organic phosphorus (DOP). The majority of natural MPn is found associated with high molecular weight dissolved organic matter and ~77% of the DOP in the Sargasso Sea exists as high molecular weight DOP (HMWDOP) (Sosa et al., 2020). MPn specifically accounts for 6.6-16.4% of this HMWDOP pool (Repeta et al., 2016; Sosa et al., 2020). Additionally, the North Atlantic has a C:P ratio of ~290-390:1 (Kolowith et al., 2001; Sosa et al., 2020). Assuming these relationships are temporally and regionally persistent, reported data for this area would imply a mean concentration of 13.7 nM MPn for surface water in the region. Sosa et al. (2020) saw ~10.7 nM MPn in the northwest Sargasso Sea (~66 nM HMWDOP * 0.165 nM MPn per nM HMWDOP), which provides confidence to our estimates. To ensure the signal was detectable using our methods, an MPn amendment concentration of 1,420 nM (~100 times greater) was used to connect all trials. Amendments of N and P macronutrients were added in ratios relative to the MPn amendment concentration (0.5:1, 1:1, 2:1, etc.). 

An additional 30 mL of ambient lab air was injected into each bottle to minimize the generation of a vacuum from time series headspace sampling. Each bottle was analyzed for the initial methane concentration in the headspace. Headspace sampling was performed in small batches to minimize disruption of incubations and methane concentrations were measured via Shimadzu GC-FID 8A with N2 as the carrier gas (Kinnaman et al., 2007). A 3 mL headspace volume was taken from each bottle by syringe and injected into the GC-FID’s gas injection manifold. The syringe was purged three times with N2 gas before headspace sample collection and between each following measurement. Data was recorded with a Shimadzu C-R68A, and the instrument was calibrated with a 3-point calibration curve at the beginning of each sampling period. Air for headspace replenishment at the end of each sampling point either came from a metal flask pressurized with air of known methane content or ambient lab air that was analyzed on the GC-FID prior to injection. 

Each condition was comprised of 3-5 replicates depending on the trial. All replicate bottles per condition were placed in a clear plastic bag and incubated in the outdoor aquaria, exposed to ambient sunlight and temperature. Control conditions accompanied each experiment and included: i) bottles containing only autoclaved seawater, ii) bottles containing autoclaved seawater and 1420 nM MPn, and iii) bottles containing autoclaved seawater and ~2 g wet weight Sargassum to ensure the measured methane was due only to MPn interaction with algal holobiont. These were compared to bottles containing autoclaved seawater, ~2 g wet weight Sargassum, and a specified concentration of MPn and/or other macronutrients. Some experiments included a dark (bottles wrapped in aluminum foil containing autoclaved seawater, ~2 g wet weight Sargassum, and 1420 nM MPn) or an antibiotic (containing autoclaved seawater, 1420 nM MPn, ~2 g wet weight Sargassum, and the antibiotic mixture) treatment. Additional details about each condition are provided in the dataset.

Experiments were also conducted with nine other species of macroalgae, five of which were collected from Bermuda and treated as described above. The other four species were collected by divers from the Pacific Ocean off the coast of Santa Barbara, California, in December 2017. Pacific incubations were conducted at the University of California, Santa Barbara in indoor aquaria equipped with UV grow lamps on a 12-hr timer that provided approximately full range of solar irradiance and were connected to flowing seawater piped in from nearshore. The experimental design followed that of the Sargassum incubations. However, the morphological features of the Pacific macroalgae required that a ½” diameter circular core was taken from the leaves of each species and placed in 150 mL borosilicate serum bottles. These headspace samples were quantified on a Picarro G2132-i Cavity Ring-Down Spectrometer (CRDS) with a slow, steady stream of carrier gas from a compressed cylinder of Breathing Air through the input port of the CRDS with an aluminum tube fitted with an injection port. Methane concentration data was collected for the carrier gas for at least 1 hour prior to each trial to establish a baseline methane concentration reading. This baseline value was adjusted accordingly for each sampling time point by taking the mean methane concentration value of the breathing air for at least 2 minutes at the beginning and ending of the sampling session. Subsamples of the bottle headspace were then injected in line with the breathing gas, allowing the methane concentration peak to fully return to baseline before the next bottle sample was injected. The times from the start and end of each bottle sample peak were recorded, and the full sampling session data file was downloaded at the end. The individual bottle peaks were then manually separated from the full file. The peak area was calculated by summing the mean methane concentration between each recorded measurement subtracted from the baseline methane concentration. Peak areas were then converted into moles of methane in the headspace by three-point calibration with methane standards.

All methane data was normalized to initial algal wet weight and headspace volume for comparison across bottles. To eliminate inclusion of extraneous methane sources in the MPn consumption signal, methane production is reported in molar quantities in excess of the mean value of control conditions containing only sterile seawater for each sampling time. All methane production rates were calculated using data collected during the first 3 days of incubation only.

References for MPn Concentration Justifications

These outlined assumptions allow the estimated natural concentration of MPn in surface Sea Water for the Atlantic Ocean, which then informs the amount of MPn in surface seawater supplied to bottles in the experiment. 

Assumptions Used to Estimate Surface Methylphosphonate Concentration (nM) in the Sargasso Sea

• A (Repeta et al., 2016; Sosa et al., 2020): HMWDOP * (0.066 to 0.164) = MPn
• B (Sosa et al., 2020): DOP * 0.77 = HMWDOP
• C (Lomas et al., 2009): TDP * 0.8 = DOP
• D (Wu et al., 2000): DOP = 0.94-0.99 * TDP
• E (Kolowith et al., 2001; Sosa et al., 2020): 290-390 C : 1 P

Note: Assumption D is used for the reported values from that citation only. Preference was given to Assumption C as it was more recently determined and thus more indicative of current conditions in the Sargasso Sea.

Reference, Nutrient Concentrations, Applied Assumptions, and Estimated Surface MPn Concentration (nM) in the Sargasso Sea

McCarthy, Hedges, & Benner, 1996   

• Nutrient Concentration: 16 mM HMWDOC                       
• Applied Assumptions: A, E                     
• Estimated Surface MPn Concentration (nM): 3.6-9.1

Canellas et al., 2000 

• Nutrient Concentration: 0-0.19 µM DOP                       
• Applied Assumptions: A, B                     
• Estimated Surface MPn Concentration (nM): 0.0-24.0

Wu et al., 2000

• Nutrient Concentration: 75±42 µM TDP                       
• Applied Assumptions: A, B, D                      
• Estimated Surface MPn Concentration (nM): 1.6-14.8

Cavender-Bares et al., 2001

• Nutrient Concentration:  0.1-0.5 µM DOP                      
• Applied Assumptions: A, B                      
• Estimated Surface MPn Concentration (nM): 5.1-63.1

Kolowith et al., 2001

• Nutrient Concentration:  42 nM HMWDOP                     
• Applied Assumptions: A                      
• Estimated Surface MPn Concentration (nM): 2.8-6.8

Mahaffey et al., 2004

• Nutrient Concentration:  0.07-0.43 µM DOP                      
• Applied Assumptions: A, B                      
• Estimated Surface MPn Concentration (nM): 3.6-54.3

Mather et al., 2008

• Nutrient Concentration:  0.03-0.31 µM DOP                      
• Applied Assumptions: A, B                      
• Estimated Surface MPn Concentration (nM): 1.5-39.2

Lomas et al., 2009

• Nutrient Concentration:  0.04-0.07 µM DOP                      
• Applied Assumptions: A, B                      
• Estimated Surface MPn Concentration (nM): 2-8.8

McLaughlin et al., 2013

• Nutrient Concentration:  0.1-0.17 µM TDP                      
• Applied Assumptions: A, B, C                    
• Estimated Surface MPn Concentration (nM): 4.1-17.2

Sosa et al., 2020                                               

• Nutrient Concentration: 66 nM HMWDOP                     
• Applied Assumptions: A                  
• Estimated Surface MPn Concentration (nM): 4.3-10.7