Independent iron and light limitation in a low-light-adapted Prochlorococcus from the deep chlorophyll maximum

Throughout the open ocean, a minimum in dissolved iron concentration (dFe) overlaps with the deep chlorophyll maximum (DCM), which marks the lower limit of the euphotic zone. Maximizing light capture in these dim waters is expected to require upregulation of Fe-bearing photosystems, further depleting dFe and possibly leading to co-limitation by both iron and light. However, this effect has not been quantified for important phytoplankton groups like Prochlorococcus, which contributes most of the productivity in the oligotrophic DCM. Here, we present culture experiments with Prochlorococcus strain MIT1214, a member of the Low Light 1 ecotype isolated from the DCM in the North Pacific subtropical gyre. Under a matrix of iron and irradiance matching those found at the DCM, the ratio of Fe to carbon in Prochlorococcus MIT1214 cells ranged from 10–40 × 10−6 mol Fe:mol C and increased with light intensity and growth rate. These results challenge theoretical models predicting highest Fe:C at lowest light intensity, and are best explained by a large photosynthetic Fe demand that is not downregulated at higher light. To sustain primary production in the DCM with the rigid Fe requirements of low-light-adapted Prochlorococcus, dFe must be recycled rapidly and at high efficiency.

S2 overnight in a 5% citranox solution, rinsing with ultrapure (18.2 MW) water, soaking in 10% hydrochloric acid for 1 week, and finally rinsed several times (>5x) in ultrapure water. Media was prepared with a seawater base obtained from 0.2 µm filtered water collected in the North Pacific subtropical gyre. Seawater was poured into 2L polycarbonate bottles (Nalgene) and microwave sterilized. Ammonium and phosphate sources were decreased from published recipes to 50 and 3.1 µM respectively. Macronutrient stocks were cleaned of contaminants using Chelex-100 resin (Bio-Rad) and all nutrient stocks were filter sterilized. Ethylene-diaminetetraacetic acid (EDTA) was added to a final concentration of 11.7 µM. Concentrations of inorganic iron (Fe') in equilibrium with EDTA were calculated as described in Sunda et al. (4), at a ratio of 10 -1.94 relative to total Fe in the media. Added Mn, Co, Ni, and Zn concentrations were 20, 10, 8, and 10 nM respectively. Background Fe concentrations were measured to be 1.5 nM in background seawater by inductively coupled mass spectrometry (ICP-MS) following extraction using a SeaFAST preconcentration system (Elemental Scientific). Additional Fe was added as a 10 mM HCl solution and media was allowed to equilibrate overnight before cultures were inoculated.
Preliminary growth experiments at a given Fe concentration were conducted in 50 mL polycarbonate bottles. 5-10 mL of exponential phase cells were then inoculated into 500 mL bottles and growth was monitored by in vivo chlorophyll fluorescence at regular (48 hr) intervals.
Approximately 4 mL of culture was poured into glass test tubes underneath a HEPA filter and then measured with a 10-AU fluorometer (Turner Designs). Specific growth rates were determined by the rate of increase in the natural log of in vivo chlorophyll fluorescence. Cultures were harvested during mid/late exponential growth.

Particulate carbon and chlorophyll analyses
Duplicate 30-50 mL aliquots of culture were filtered onto 25 mm GF/F filters (Whatman). For chlorophyll analyses, frozen filters were extracted in 90% acetone overnight at -20ºC and extracted chlorophyll was quantified on a 10-AU fluorometer. Prior to use, filters for particulate carbon were combusted at 500 ºC for >3 hours. After filtration, filters were dried over 48 hours at 60 ºC and then measured with a 4010 Elemental Analyzer (Costech), calibrated with methionine and acetanilide standards, as described previously (5).

S3
For particulate metal analyses, 100-200 mL aliquots were filtered onto 47 mm 0.2 µm polyethersulfone membranes under vacuum in a class 1000 clean room. Filters were digested overnight at 95 ºC in 30 mL perfluoroalkoxy vials (Savillex) by refluxing 5 mL of 50% distilled nitric acid (HNO3) with 1 ppb Indium added. Filters were then removed and samples were dried down at 100ºC and subsequently digested in 200 µL of 1:1 HNO3:HCl for 2 hours. Digested metals were dried down again and re-dissolved in 0.1 M HNO3 for analysis on an Element2 ICPMS (Thermo Fisher). Fe concentrations were determined in medium resolution mode relative to a standard curve of 0.1-100 ppb, diluted from a certified reference (Inorganic Ventures).
Intensity of the 115 In peak was used to correct for matrix effects, as well as sample loss during digestion (e.g. acid adsorbed onto the filter). Combined reagent and filter blanks (42 ± 21 pmol, n = 10) were quantified and subtracted from sample values (180 to 3,200 pmol). Fe concentrations were calculated after dividing by the volume filtered, which was determined gravimetrically.
We note that filters were not washed with chelating solutions prior to digestion. Initial experiments suggested a significant amount of cell lysis occurred during short (~5 min) rinses in an oxalate-EDTA solution. As a result, particulate Fe concentrations may contain some amount of Fe bound to the cell surface. Most of our experiments are well below the Fe hydroxide precipitation thresholds described by Sunda and Huntsman (500 pM Fe' (6)), meaning that this effect is minor, especially when growth was Fe-limited.

Comparisons to HOT data
Primary production measurements for Hawaii Ocean Time-series cruises 1-298 were accessed via the HOT-DOGS website (hahana.soest.hawaii.edu/hot/hot-dogs/interface.html; ref. 7). Measurements made at 100 m (n = 273) were averaged. All dissolved iron data between 90 to 120 m from Fitzsimmons et al. (8) were averaged. It should be noted that there is a summertime bias in the dFe measurements at Station ALOHA, especially in this depth range. Because primary production at 100 m in summer will be on the higher end of the range quoted in text, the resulting turnover times may be on the lower end of the quoted range.  (12)) Chl per PSU 272 = 570/3 + 164/2 Chl per cell 1.82 x 10 6 cell -1 = 272 x 6700 Chl:C ratio 760 x 10 -6 mol mol -1 = 1.82 x 10 6 / 2.41 x 10 9 Measured Chl a:C (Fe replete) 330 x 10 -6 mol mol -1 Mean of 15 and 50 nM Fe treatments, Table S1 Chl a:b ratio 0.6-1.2 mol mol -1 (Moore et al. 1997 (16)) Estimated Chl:C (Fe replete) 530-730 x 10 -6 mol mol -1 = 330 x 1.6, 330 x 2.2