Distributions of particulate Heme b in the Atlantic and Southern Oceans—Implications for electron transport in phytoplankton


  • Martha Gledhill,

    Corresponding author
    1. School of Ocean and Earth Science, National Oceanography Centre, University of Southampton, Southampton, UK
    2. Chemical Oceanography, Helmholtz Centre for Ocean Research (GEOMAR), Kiel, Germany
    • Corresponding author: M. Gledhill, Chemical Oceanography, Helmholtz Centre for Ocean Research (GEOMAR), DE-24148 Kiel, Germany. (mgledhill@geomar.de)

    Search for more papers by this author
  • Eric P. Achterberg,

    1. School of Ocean and Earth Science, National Oceanography Centre, University of Southampton, Southampton, UK
    2. Chemical Oceanography, Helmholtz Centre for Ocean Research (GEOMAR), Kiel, Germany
    Search for more papers by this author
  • David J. Honey,

    1. School of Ocean and Earth Science, National Oceanography Centre, University of Southampton, Southampton, UK
    Search for more papers by this author
  • Maria C. Nielsdottir,

    1. School of Ocean and Earth Science, National Oceanography Centre, University of Southampton, Southampton, UK
    2. Department of Ocean, Earth, and Atmospheric Sciences, Old Dominion University, Norfolk, Virginia, USA
    Search for more papers by this author
  • Micha J. A. Rijkenberg

    1. School of Ocean and Earth Science, National Oceanography Centre, University of Southampton, Southampton, UK
    2. Royal Netherlands Institute for Sea Research, Den Burg, Netherlands
    Search for more papers by this author


[1] Concentrations of heme b, the iron-containing component of b-type hemoproteins, ranged from < 0.4 to 5.3 pM with an average of 1.18 ± 0.8 pM (± 1σ; n = 86) in the Iceland Basin (IB), from < 0.4 to 19.1 pM with an average of 2.24 ± 1.67 pM (n = 269) in the tropical northeast Atlantic (TNA) and from 0.6 to 21 pM with an average of 5.1 ± 4.8 pM (n = 34) in the Scotia Sea (SS). Heme b concentrations were enhanced in the photic zone and decreased with depth. Heme b concentrations correlated positively with chlorophyll a (chl a) in the TNA (r = 0.41, p < 0.01, n = 269). Heme b did not correlate with chl a in the IB or SS. In the IB and SS, stations with high-chlorophyll and low-nutrient (Fe and/or Si) concentrations exhibited low heme b concentrations relative to particulate organic carbon (< 0.1 μmol mol−1), and high chl a:heme b ratios (> 500). High chl a:heme b ratios resulted from relative decreases in heme b, suggesting proteins such as cytochrome b6f, the core complex of photosystem II, and eukaryotic nitrate reductase were depleted relative to proteins containing chlorophyll such as the eukaryotic light-harvesting antenna. Relative variations in heme b, particulate organic carbon, and chl a can thus be indicative of a physiological response of the phytoplankton community to the prevailing growth conditions, within the context of large-scale changes in phytoplankton community composition.

1 Introduction

1.1 Heme b

[2] Oceanic phytoplankton are known to change the abundance of iron-containing proteins when nutrient concentrations or incident light levels change [Pankowski and McMinn, 2009; Peers and Price, 2006; Saito et al., 2011; Strzepek and Harrison, 2004; Wolfe-Simon et al., 2006]. The cellular abundance of such proteins influences the efficiency of photosynthesis [Bailey et al., 2008; Cardol et al., 2008; Peers and Price, 2006] and so links nutrient abundance and light to ocean productivity. Ferredoxin [Mckay et al., 1999; Pankowski and McMinn, 2009], photosystem I (PSI), cytochrome b6f (cytb6f) [Eberhard et al., 2008; Saito et al., 2011; Strzepek and Harrison, 2004], cytochrome c550 [Saito et al., 2011], soluble cytochrome c [Eberhard et al., 2008; Peers and Price, 2006], and iron superoxide dismutase [Wolfe-Simon et al., 2006] have all been shown to be reduced in abundance or replaced by noniron-containing proteins when oceanic phytoplankton are grown under low-iron or -nutrient conditions in the laboratory. Furthermore, recent work has shown downregulation of genes associated with iron proteins or iron protein synthesis [Allen et al., 2008; Lommer et al., 2012; Thompson et al., 2011] and diurnal cycling of iron protein pools in order to conserve iron use [Saito et al., 2011]. However, few studies have investigated the abundance of iron-containing proteins in the field or how their distributions are influenced by the prevailing nutrient and light regimes [Erdner and Anderson, 1999; Pankowski and McMinn, 2009]. Heme b is an iron-containing porphyrin which functions as a prosthetic group for proteins involved in electron transfer and the scavenging of reactive oxygen species [Chapman et al., 1997]. Heme b is highly toxic to cells if not incorporated into proteins [Espinas et al., 2012] and the stoichiometry of heme b within proteins is fixed [Shekhawat and Verma, 2010; Tanaka and Tanaka, 2007], so that changes in heme b abundance will reflect changes in the cellular abundance of heme b-containing proteins. In marine phytoplankton, heme b is incorporated into proteins involved in photosynthesis, respiration, and reactive oxygen scavenging [Honey et al., 2013]. Laboratory studies of marine eukaryotic phytoplankton showed that heme b concentrations extracted by ammoniacal detergent averaged 3.7 ± 0.9 μmol mol−1 C in resource replete conditions and that the cellular concentrations of heme b decreased under nutrient deplete conditions. Heme b accounted for 18 ± 14% of the total particulate iron pool when iron concentrations in the culture media were low. Decreases in heme b relative to particulate organic carbon (POC) were also observed to correspond with decreasing nitrate and phosphate concentrations across a transect of the (sub-) tropical North Atlantic. Chlorophyll a:heme b ratios were also found to vary for certain species when phytoplankton were grown under decreased iron, nitrate, or light conditions. However, results from the temperate and (sub-) tropical North Atlantic showed that chl a:heme b increased with depth and were thus driven by gradients in light rather than nutrients [Honey et al., 2013]. Determination of the concentration of heme b in particulate material and comparison with phytoplankton biomass, POC, and chl a in the ocean could thus provide information on the variation in heme b protein abundance and the response of phytoplankton to prevailing nutrient and light conditions.

[3] In this study, we determined heme b in particulate material in three contrasting regions of the Atlantic Ocean and Southern Ocean: the Iceland Basin (IB), the tropical northeast Atlantic (TNA), and the Scotia Sea (SS). The IB is situated in the high-latitude North Atlantic and is characterized by pronounced diatom-dominated spring blooms [Leblanc et al., 2009], which nevertheless do not result in the complete drawdown of nitrate and phosphate. Residual nutrient stocks are thus subsequently exploited by more mixed phytoplankton assemblages [Leblanc et al., 2009; Poulton et al., 2010]. The post spring bloom period is iron limited [Nielsdottir et al., 2009], a result of both low atmospheric iron inputs [Jickells et al., 2005] and suboptimal iron:nitrate ratios in winter-overturned deep waters [Nielsdottir et al., 2009].

[4] The region of the TNA examined in this study is influenced in the east by the highly productive upwelling system off the Northwest African Coast, in the south by the oxygen minimum zone off the Cape Verde Islands and in the northwest by the permanently stratified oligotrophic subtropical North Atlantic Gyre [Stramma et al., 2008]. Atmospheric iron inputs in this region are relatively high due to the influx of Saharan and Sahel dust [Jickells et al., 2005; Mulitza et al., 2010; Stuut et al., 2005] with maximum dust inputs in this region occurring in winter (January to March) [Chiapello et al., 1995]. The phytoplankton community in the TNA is dominated by picoplankton comprising Prochlorococcus sp., Synechococcus sp., and picoeukaryotes [Hill et al., 2010], although nitrogen-fixing Trichodesmium spp. are also observed in this region [Moore et al., 2009], with their abundance being tightly linked to dust inputs [Rijkenberg et al., 2011, 2012].

[5] The SS lies within the Atlantic sector of the Southern Ocean. The SS contrasts with much of the high-nutrient low-chlorophyll (HNLC) regions of the Southern Ocean as it supports extensive phytoplankton blooms, with the South Orkney Islands, the South Sandwich Islands, and South Georgia acting as potential sources of iron via the “island mass effect” [Doty and Oguri, 1956; Korb et al., 2005; Nielsdottir et al., 2012]. The waters around South Georgia are particularly productive and can support Antarctic diatom-dominated blooms of up to 12 mg m−3 chl a for periods of up to 5 months [Korb et al., 2005]. However, productivity in the area is patchy, and regions to the south of South Georgia have HNLC characteristics [Hinz et al., 2012; Korb et al., 2010].

[6] We compare the distribution of heme b to chl a, phytoplankton biomass, POC, and particulate organic nitrogen (PON) and interpret our findings within the context of the prevailing nutrient (dissolved iron (dFe), nitrate, phosphate, and silicate) and light environments. We assess how the relative abundance of heme b varies in the ocean, where chl a:heme b ratios increase and how changes in the relative abundance of heme b proteins and light-harvesting complexes relate to nutrient distributions.

2 Methods

2.1 General

[7] Samples were obtained during three cruises in 2007 and 2008 (Figure 1 and Table 1). The IB cruise took place from 25 July to 19 August 2007. The cruise consisted of three repeat surveys of a 90 × 90 km grid in the North Atlantic subpolar gyre. The TNA cruise took place from 6 January to 3 February 2008. Samples were collected in the vicinity of the Canary and Cape Verde Islands, the subtropical North Atlantic Gyre and on a latitudinal transect along 12°N. The SS cruise took place from 1 January to 11 February 2008. Samples were collected in the vicinity of the South Orkney Islands, within the Antarctic Circumpolar Current (ACC) and in the bloom associated with South Georgia. Hydrographic data were collected with a Seabird 9/11 + conductivity-temperature-depth (CTD) on the IB and TNA cruises and with a Seabird SBE9 + CTD on the SS cruise.

Figure 1.

Map of the Atlantic Ocean showing the areas sampled. Inset maps show the sampled stations superimposed on average Sea-viewing Wide Field-of-view Sensor-derived chl a (mg m−3) for the times of the cruises. (a) Iceland Basin, (b) tropical northeast Atlantic, and (c) Scotia Sea. Note that scale changes for chl a. Some station numbers are omitted in Figure 1b for clarity.

Table 1. Heme b Extraction Efficiencies From Representative Hemoproteins Using Ammoniacal Detergenta
HemoproteinRecovery (%)Approximate Distance Between Iron Atom and Protein Surface (Å)d5th/6th Iron LigandbSpin Stateb
  1. a

    An estimate of exposure of the heme group to the solvent, expressed as the distance between the center of the heme group and the protein surface, the ligands occupying the fifth and sixth coordinate sites in the octahedrally complexed iron, and the spin state of iron in the complex are also provided.

  2. b

    Frausto da Silva and Williams, 2001.

  3. c

    Honey et al., 2013.

  4. d

    Estimated using Chem3D Pro© from Protein Data Bank files for Bovine catalase (3NWL) [Foroughi et al., 2011], horseradish peroxidase (1HCH) [Berglund et al., 2002], hemoglobin (1GZX) [Paoli et al., 1996], myoglobin (2V1K) [Hersleth et al., 2007], and rat cytochrome b5 (1MNY) [Banci et al., 2001].

Catalase8 ± 1c>20Tyrosine/H2OHS
Horseradish peroxidase36 ± 4c10–20Histidine/H2OHS
Hemoglobin76 ± 4<10Histidine/H2OHS/LS
Myoglobin79 ± 2<10Histidine/H2OHS/LS
Cytochrome b576 ± 1<10Histidine/HistidineLS

2.2 Sample Collection

[8] Samples for heme b, chl a, POC, PON, and major nutrients were collected from depths < 250 m using a stainless steel CTD rosette equipped with 20 L Niskin (General Oceanics) bottles. Samples for dFe were collected using a trace metal clean titanium CTD rosette equipped with 12 L trace metal clean Teflon-coated OTE bottles or using a trace metal clean fish towed at approximately 5 m depth while the ship was steaming at 10 knots [Nielsdottir et al., 2012, 2009].

2.3 Particulate Heme b

[9] Between 1 and 4 L of seawater was filtered over 0.7 μm glass fiber filters (MF300, Fisher) and frozen (< −80°C) for analysis in the laboratory. An operationally defined fraction of the total intracellular heme b pool [Honey et al., 2013] was extracted into 1 mL of 0.02 M NH4OH containing the zwitterionic detergent EMPIGEN (2.5% v/v, Sigma). Extracts were centrifuged, filtered (0.2 μm, Minisart, Sartorius), and determined spectrophotometrically after separation from other pigments by high-performance liquid chromatography [Gledhill, 2007; Honey et al., 2013]. Separations were performed using a polystyrene divinyl benzene stationary phase (PLRP-S column, 50 × 2.1 mm, 5 μm, Varian Inc.). High-performance liquid chromatography was carried out using binary gradient high-pressure pumps (Shimadzu, LC-10ADVPμ) and a diode array spectrophotometer equipped with a microcell (Shimadzu, SPD-M10AVP). The 100 μL of sample was injected onto the column using an autosampler (Shimadzu, SIL-10ADVP). The system was computer operated (LCsolution software) via a controller (Shimadzu, SCL-10AVP). Mobile phases consisted of (A) 20:20:40:0.1% (v:v:v:v) isopropanol (IPA):acetonitrile (ACN):water:nonafluoropentanoic acid (NFPA), and (B) 1:1:0.1 (v:v:v) IPA:ACN:NFPA. A standard gradient of 60% A to 100% B over 15 min, followed by 5 min isocratic elution with 100% B was used. The flow rate was 200 μL min−1. Elution of heme b (retention time 5.5 min) was monitored by absorbance at 400 nm after applying a background absorbance correction using 450 nm in order to remove baseline fluctuations caused by EMPIGEN. Chromatograms and UV-visible light spectrometer spectra were routinely examined for the presence of other potential interfering pigments, which can be distinguished by shifts in the maximum absorption wavelength and the presence of secondary absorbance maxima. Heme b was quantified by comparison with standard solutions of iron (III) protoporphyrin IX chloride (Frontier Scientific). The detection limit for the technique was 1.6 nM. Extraction efficiency of heme b from human hemoglobin, horse skeletal myoglobin and human cytochrome b5 (Sigma) proteins using ammoniacal EMPIGEN were assessed as described previously [Honey et al., 2013]. The extracted concentration of cytochrome b5 was compared to the concentration determined from the difference spectrum for the reduced versus oxidized protein.

2.4 Chl a, Fv/Fm POC/PON, and Nutrients

[10] Water samples (100–200 mL) for chl a analysis were filtered over glass fiber filters (MF300, Fisher). Chlorophyll a was extracted into 10 mL of 90% (v/v) acetone over a 20–24 h period in the dark (4 °C). Chlorophyll a was determined by fluorometry (Turner Design TD-700) and calibrated using a pure chl a standard (spinach, Sigma). The maximum quantum yield (Fv/Fm) is reported here for dawn stations in the IB (IB222, IB226, IB260, and IB285) and SS. Fv/Fm was determined via chlorophyll fluorescence measurements obtained using fast repetition rate fluorometry (FRRF, Chelsea Scientific Instruments) and Fluorescence Induction and Relaxation (FIRe, Satlantic) [Bibby et al., 2008; Kolber et al., 1998]. Particulate organic carbon and PON were determined on the IB cruise following filtration of 1–2 L seawater onto precombusted glass fiber (0.7 μm; Whatman) filters [Poulton et al., 2006]. On the TNA cruise, POC and PON were determined at selected stations by mass spectrometry after filtration of up to 4 L seawater [Moore et al., 2006]. Macronutrients (nitrate + nitrite, hereafter termed nitrate, phosphate, and orthosilicic acid) were determined on board using an autoanalyzer and standard colorimetric techniques [Sanders and Jickells, 2000; Whitehouse, 1997] or in the TNA, with a nanomolar nutrient system following the method of Patey et al. [2008].

2.5 Dissolved Iron

[11] Dissolved iron was analyzed after gentle pressure filtration (0.2 μm) by an automated flow injection chemiluminescence method as previously described [Nielsdottir et al., 2012, 2009; Obata et al., 1993].

3 Results and Discussion

3.1 Heme b Extraction

[12] We have previously shown that ammoniacal detergent extraction of heme b is operational defined [Honey et al., 2013]. Comparison with total cellular iron indicated recoveries of 18 ± 14% of iron as heme b, consistent with the theoretical percentage of heme b in the photosystem proteins [Gledhill, 2007]. However, recoveries of heme b from catalase and horse radish peroxidase indicated that heme b extraction was below 40% for these hemoproteins. In this study, in order to better define ammoniacal detergent heme b extractions, we determined extraction efficiencies for three further readily available hemoproteins, hemoglobin, myoglobin, and mitochondrial cytochrome b5. The results showed recoveries > 75–80% for extraction of heme b from hemoproteins such as globins and cytochromes using ammoniacal detergent (Table 1). This is likely related to the heme b ligand environment, spin state, and the degree to which the heme group is likely to be exposed to the solvent (Table 1) and is consistent with the known lability of heme b in hemoproteins, previously reported heme b to cellular iron ratios and the observed relationship between heme b and the photosystem II protein PsbA [Frausto da Silva and Williams, 2001; Honey et al., 2013; Weber et al., 2011]. Therefore, heme b concentrations determined using our method are likely to be strongly influenced by the abundance of b-type cytochromes, which in phytoplankton include cytb6f, cytochrome b5, and cytochrome bc1. In addition, it is possible that the intracellular location of heme will also play a role in extraction efficiency. Thus, although not fully quantitative, variation in our heme b concentrations will reflect variation in the cytochrome content and thus changes in the potential for electron transport within phytoplankton populations.

3.2 Surface Heme b, Chl a, POC, and PON in the Iceland Basin

[13] Particulate heme b concentrations in the IB ranged from < 0.4 to 5.3 pM, averaging 1.18 ± 0.8 pM (± 1 standard deviation; n = 86). Depth profiles showed some enhanced concentrations in surface waters, but trends were not as strong as for chl a (Figure 2). Integrated mixed-layer depth values for chl a, heme b, POC, and PON calculated for the IB are given in Table 2. Heme b concentrations in the IB were found to correlate weakly with POC and PON; however, no significant correlation was observed with chl a (Table 3). Comparison of mixed-layer depth-integrated heme b concentrations with phytoplankton carbon (phyC) [Poulton et al., 2010] indicated that heme b:phyC ratios were between 0.07 and 0.78 μmol mol−1, considerably lower than ratios observed in phytoplankton grown under nutrient replete conditions in the laboratory (average 3.7 ± 0.9 μmol mol−1) and lower than those reported previously for the Celtic Sea [Honey et al., 2013]. The low concentration of heme b in this region was also reflected in the heme b:POC and heme b:PON ratios (Table 3). The low concentrations of heme b with respect to carbon indicate that the abundance of b-type cytochromes was also low and is compatible with the low Fv/Fm values (Table 4) observed in the mixed layer in this region at the time of sampling. Phytoplankton have been found to decrease their heme b:POC ratios in response to nutrient limitation [Honey et al., 2013]. At station IB285, situated in the center of a cyclonic eddy, low heme b:POC ratios (0.06 ± 0.01 μmol mol−1) were not reflected in decreases in chl a:POC (~ 80 μmol mol−1), which were close to the average obtained for the IB (107 ± 24 μmol mol−1). The decrease in heme b:POC thus led to an integrated mixed-layer chl a:heme b ratio of 1260 at this station. Integrated chl a:heme b ratios > 500 were also observed at stations IB260 and IB274, and again, coincided with low heme b:POC ratios (0.07 ± 0.01 and 0.09 ± 0.03 μmol mol−1, respectively). Coccolithophores, dominated by Emiliania huxleyi made up between 10 and 25% of the phytoplankton carbon at the time of this study [Poulton et al., 2010]. Chl a to heme b ratios in E. huxleyi have been observed to increase under low-iron and/or -light conditions in the laboratory [Honey et al., 2013]. Average daily irradiances in the IB during the cruise ranged between 9 and 39 mol photosynthetically active radiation m−2 d−1 and were not thought to be limiting phytoplankton growth in the mixed layer [Poulton et al., 2010]. The concentrations of nitrate and phosphate at stations IB260, IB274, and within the cyclonic eddy were also not fully depleted (average mixed-layer depth concentrations at these stations: 3.6 ± 2.1 μM nitrate, 0.29 ± 0.1 μM phosphate, 0.5 ± 0.4 μM silicate). Mixed-layer depth dFe concentrations averaged 0.016 ± 0.01 nM at the cyclonic eddy station and 0.05 ± 0.01 nM at station IB260 (no dFe determined at station IB274), constituting the lowest dFe concentrations detected during the IB study. The high chl a:heme b ratios observed in the IB are therefore most likely a response of the phytoplankton population to the low dFe concentrations observed at stations IB260, IB274, and in the cyclonic eddy.

Figure 2.

Vertical profiles for heme b (red triangles), chl a (green squares), and chl a:heme b (open circles) determined at stations in the Iceland Basin. Error bars for heme b represent the concentration range determined for each sample (n = 2).

Table 2. Integrated Mixed-Layer Depth (MLD) Values for Heme b (nmol m−2), Chlorophyll a (μmol m−2), POC (mmol m−2), and PON (mmol m−2) Concentrations Calculated for Stations Sampled in the Iceland Basin, Tropical North Atlantic, and Scotia Sea
StationLatitude (°N)Longitude (°W)Julian DayMLDIntegrated Heme b (nmol m−2)Integrated Chl a (μmol m−2)Integrated POC (mmol m−2)Integrated PON (mmol m−2)
Iceland Basin
Tropical Northeast Atlantic
Scotia Sea
Table 3. Results of Pearson Product Moment Correlation for Heme b, Chl a, Particulate Organic Carbon (POC), and Particulate Organic Nitrogen (PON) in the Iceland Basin, Tropical Northeast Atlantic, and Scotia Seaa
  1. a

    No POC/PON data were collected for the Scotia Sea.

Iceland Basin
Heme b r (n)0.19 (84)0.41 (77)0.45 (77)
Chl a r (n) 0.66 (79)0.6 (79)
p <0.01<0.01
POC r (n)  0.99 (79)
p  <0.01
Tropical Northeast Atlantic
Heme b r (n)0.41 (269)0.23 (60) 
Chl a r (n) 0.19 (60)0.38 (60)
p 0.15<0.01
POC r (n)  0.60 (64)
p  <0.01
Scotia Sea
Heme b r (n)0.27 (34)  
Table 4. Comparison of Heme b:PhyC Ratio With Fv/Fm Observed in the Iceland Basin and Scotia Sea
 Heme b:PhyC (μmol mol−1)Fv/Fm
  1. a

    Integrated mixed-layer heme b:phyC ratios are compared with average mixed-layer Fv/Fm values.

  2. b

    Values for heme b:phyC ratios and Fv/Fm for samples collected at 20 m.

Iceland Basina
IB2220.240.27 ± 0.02
IB2260.330.27 ± 0.02
IB2600.040.24 ± 0.01
IB2850.090.29 ± 0.01
Scotia Seab

3.3 Surface Heme b, Chl a, POC, and PON in the Tropical North Atlantic

[14] Particulate heme b in the TNA averaged 2.24 ± 1.67 pM (n = 275) and were very similar to those observed in 2010 in this region [Honey et al., 2013]. Minimum concentrations were below the detection limit (< 0.4 pM), while the maximum concentration observed in this region was 19.4 pM. Depth profiles were similar to those of chl a, with enhanced concentrations in the photic zone (Figure 3). In the TNA, there was a weak correlation between heme b and PON, but no significant relationship between heme b and POC (Table 3). Integrated mixed-layer depth values for chl a, heme b, POC, and PON calculated for the TNA are given in Table 2. Mixed-layer heme b:POC and heme b:PON ratios were higher (1.0 ± 0.5 and 7.6 ± 2.9 μmol mol−1, respectively) than values observed for the IB, although still lower than heme b:POC ratios observed for nutrient replete phytoplankton in the laboratory [Honey et al., 2013]. There was no overall change in heme b:POC with depth (data not shown), indicating that light was unlikely to be influencing the level of heme b in this region, despite the presence of a deep chlorophyll maximum at some stations in the TNA study area. Chl a to POC ratios were also higher (average 134 ± 80 μmol mol−1) in the TNA compared to the IB (Table 2). A significant correlation between heme b and chl a was observed in the TNA samples (Table 3) consistent with previously reported results [Honey et al., 2013]. Chl a to heme b ratios in the TNA were below 500, except for four isolated samples (Figure 3c). The frequency of samples at depth with high chl a:heme b ratios makes it difficult to identify a cause, although the heme b concentrations detected in these samples were likely to be subject to greater error as they were close to the detection limit of the heme b technique. Integrated mixed-layer depth chl a:heme b ratios averaged 124 ± 45, similar to ratios reported previously for the region [Honey et al., 2013] and considerably lower than those observed in the IB. Nitrate and phosphate concentrations were very low in surface waters in the TNA (3–50 nM nitrate, 2–90 nM phosphate) and are likely to limit growth in the region [Moore et al., 2009]. A correlation between these limiting nutrients and heme b has been reported for a transect across the (sub-) tropical North Atlantic Ocean [Honey et al., 2013]; however, in this study no significant correlation was observed, likely a result of the reduced spatial scale of the survey.

Figure 3.

Contour plot for vertical sections of (a) chl a concentration (nM), (b) heme b concentration (pM), and (c) chl a:heme b determined in the tropical North Atlantic. Station numbers are indicated on Figure 3b.

[15] On 14 days in the TNA, samples were collected both at dawn and midafternoon (Table 1). Although these samples were not collected from the same waters, afternoon (1500–1700 GMT) mixed-layer depth chl a:heme b ratios were higher than dawn (0700 GMT) ratios on 11 of the 14 days (Figure 4). These results indicate that diurnal variation in the production of chl a and heme b may be occurring, with heme b production lower during the day relative to chl a as a result of optimization of the chl a/heme b biosynthetic pathway [Papenbrock et al., 1999; Tanaka and Tanaka, 2007]. The lack of concurrent POC/PON data does not allow for examination of potential diurnal cycling of the overall abundance of heme b proteins, which has been reported in laboratory cultures [Saito et al., 2011].

Figure 4.

Integrated mixed-layer depth chl a:heme b ratios during the day in the tropical northeast Atlantic where stations were sampled at dawn (0700 GMT, dark circles) and midafternoon (1500–1700 GMT, open circles).

3.4 Surface Heme b and Chl a in the Scotia Sea

[16] Heme b concentrations in the SS varied from 0.6 to 21 pM with an average of 1.9 ± 1.7 pM and were thus higher than those determined in the IB (Figure 5). Particulate organic carbon and PON were not determined for the SS samples. Phytoplankton counts were undertaken at a depth of 20 m at each station [Korb et al., 2010] and heme b concentrations were therefore compared to the calculated phytoplankton carbon biomass (phyC). However, it should be noted that these calculations do not include the biomass from nanoeukaryotes or picoeukaryotes and are thus likely to overestimate actual heme b:phyC ratios. Results of the comparison are reported in Table 4. Heme b to phyC ratios observed at SSC2, SSC4, and SSP3 were similar to heme b:C ratios observed in nutrient replete phytoplankton grown in the laboratory [Honey et al., 2013]. Bioassay experiments carried out close to SSC2 and SSP3 indicated that phytoplankton were not iron limited [Hinz et al., 2012]. The low heme b:phyC ratio observed at stations SSP24, SSP28, and SSR3 to the northwest of South Georgia and in the SS indicates that the phytoplankton community was growing under low-nutrient conditions at these stations. Iron bioassay experiments near station SSP24 indicated that the phytoplankton community was iron limited in this area [Hinz et al., 2012]. The estimates of heme b:phyC determined for stations in the SS are broadly compatible with changes in Fv/Fm recorded in the region (Table 4). Fv/Fm was highest around station SSR3, where the highest heme b:phyC ratio was observed. Furthermore, south of South Georgia at stations SSP24 and SSP28, Fv/Fm was also observed to be lower (<0.3, Table 4), consistent with the depleted heme b:phyC ratios. However, at other stations (SSC2, SSC4, and SSP3) with similar Fv/Fm values, we observed a wide variability in heme b:phyC ratios. This could have been a result of underestimation of the heme b:phyC ratio due to omission of C from nanoeukaryotes and picoeukaryotes, likely to be especially important at SSC2 and SSC4, as these areas were dominated by small phytoplankton species [Korb et al., 2011]. In addition, comparison between Fv/Fm and heme b:phyC in the region is also likely influenced by changes in the community composition [Suggett et al., 2009] as the dominant taxanomic groups varied from cryptophytes at station SSC2, dinoflagellates at SSC4, and diatoms at SSP24, SSP28, SSP3, and SSR3 [Korb et al., 2010, 2011]. It was notable that chl a:phyC ratios at stations SSP24, SSP28, and SSR3 were close to the average observed for all the stations (252, 265, and 590 μmol mol−1, respectively), showing that while heme b was depleted relative to biomass at these stations, chl a was not. Integrated mixed-layer depth values for chl a and heme b calculated for the SS are given in Table 2. Chlorophyll a to heme b ratios ranged between 80 and 1940, with a mean value of 533 (Figure 5). As with the IB, no significant correlation between heme b and chl a was observed (Table 3). Integrated chl a:heme b ratios were below 500 north of the South Orkney Islands and increased going north across the SS to a value of 528 just south of South Georgia. West of South Georgia mixed-layer chl a:heme b ratios decreased to 240. This region is strongly influenced by iron inputs from the South Georgia Ridge and the station sampled was close to an area reported to be iron replete at the time of sampling (Figure 6) [Hinz et al., 2012]. Northwest of South Georgia, however, we observed high chl a:heme b ratios (> 1400), combined with low heme b:phyC ratios. This station (SSR3) was characterized by high chl a (~ 5 μg L−1), complete drawdown of Si (< 0.15 μM) and relatively depleted N concentrations (~ 14 μM). The phytoplankton community at both stations was dominated by diatoms [Korb et al., 2010, 2011]; however, Fv/Fm was lower at SSR3 than at SSP3 (Table 4). Dissolved iron concentrations were not recorded at the station itself; however, comparison of dFe and nutrient data showed that the closest underway dFe samples (situated 15 and 27 km to the southwest of the station; upstream in the ACC) had dFe concentrations of 0.16 ± 0.01 nM and 0.61 ± 0.1 nM, respectively (Figure 6). The phytoplankton community at SSR3 was dominated by centric diatoms over 10 μm in size [Korb et al., 2010] which are thought to have relatively high iron requirements [Ho et al., 2003; Sunda and Huntsman, 1995]. Therefore, it is possible that the phytoplankton at this station have exhausted the nutrient supply to the extent that they have reduced their hemoprotein content.

Figure 5.

Vertical profiles for heme b concentrations (red triangles), chl a concentrations (green squares), and chl a:heme b ratio (open circles) calculated for stations in the Scotia Sea. Error bars for heme b represent the concentration range determined for each sample (n = 2).

Figure 6.

Surface distribution of (a) dissolved Si and (b) dissolved iron northwest of South Georgia in relation to stations SSP3 (star) and SSR3 (circle).

3.5 Implications of Changes in Intracellular Heme b for Marine Phytoplankton

[17] This study reports the first investigation into the distribution of heme b in iron-limited marine environments. In the Atlantic and Southern Oceans, we have observed variations in heme b concentrations ranging from < 0.2 to 21 pM. Heme b distributions reflected changes in phytoplankton biomass, with higher heme b concentrations observed in regions of higher biomass. However, the concentration of heme b relative to the organic carbon and nitrogen and to chl a varied considerably between the three areas surveyed. Our study indicates that variations of heme b:POC ratio reflect changes in the abundance of heme b proteins such as cytb6f, PSII, and eukaryotic nitrate reductase in particulate material while variations in chl a:heme b ratio will reflect changes in the relative capacity for light harvesting versus electron transport. Laboratory studies have shown that heme b:POC decreases when marine phytoplankton are grown under low-nutrient conditions. Species specific increases in chl a:heme b ratios were also observed in phytoplankton grown under low-nutrient and low-light conditions [Honey et al., 2013]. In the IB and SS to the south of South Georgia, chl a:heme b ratios over  500 were observed at stations with low dFe concentrations (Figure 7). The high chl a:heme b ratios resulted mainly from a relative reduction in heme b abundance. Although dFe concentrations in the SS to the northwest of South Georgia were not as low as those in the IB and south of South Georgia, we suggest that the low Si concentrations coupled to the higher iron demand of the dominant large centric diatoms led to the enhanced chl a:heme b ratios. In this study, we have highlighted increases in the chl a:heme b ratio to values > 500, which are in excess of chl a:heme b ratios we have observed in the laboratory [Honey et al., 2013]. Observed increases in chl a:heme b to values > 500 were associated with particular stations (e.g., IB285 and SSR3) which were not observed to have major shifts in species composition when compared to other nearby stations [Korb et al., 2010; Poulton et al., 2010]. This suggests that the change in chl a:heme b ratio in the IB and SS resulted from a physiological response, rather than a shift in species composition. However, the absence of high chl a:heme b ratios in the TNA may not be influenced solely by the higher dFe and heme b:POC ratios in this region when compared to the SS and IB but is also likely to reflect large-scale changes in phytoplankton community composition as biomass in this region is dominated by prokaryotes which utilize different light-harvesting strategies [Ting et al., 2002].

Figure 7.

Integrated chl a:heme b ratios plotted against average mixed-layer dFe concentrations. Iceland Basin—inverted triangles, tropical northeast Atlantic—circles, and Scotia Sea—squares. For the Scotia Sea, the closest dFe concentrations collected with underway sampling system were used.

[18] The chl a:heme b ratios reported here are likely consistent with the adaptation of the photosystem to the prevailing light and nutrient conditions, within the context of large-scale changes in phytoplankton community dynamics. Our results indicate that the abundance of heme b and thus cytochromes with respect to biomass and other photosystem complexes are likely to change. Although we cannot be certain which cytochromes are most affected, a reduction in cytb6f would be particularly crucial as it connects the two photosynthetic reaction center complexes and is involved in linear and cyclic electron transfer [Baniulis et al., 2008; Iwai et al., 2010]. Cytochrome b6f proteins and genes have been shown to be downregulated in some phytoplankton when iron stressed [Strzepek and Harrison, 2004; Thompson et al., 2011]. Cytochrome b6f potentially plays a contributory role in regulating electron transport through the process of “photosynthetic control” [Eberhard et al., 2008; Peltier et al., 2010]. Furthermore, in green algae, cytb6f regulates state transitions and the switching of light-harvesting complexes from PSII to PSI thus changing the amount of linear versus cyclic electron transport [Lemeille and Rochaix, 2010]. Gaining insight into cytb6f abundance in the ocean and its relationship to other photosystem complexes will shed light on fundamental electron transport processes occurring in the ocean, and how these impact on productivity.

[19] Our investigation into heme b distributions has shown that heme b:POC and chl a:heme b ratios vary with biomass and nutrient distributions. We have identified stations in the SS and IB where phytoplankton have high chl a:heme b as a result of low-nutrient concentrations. Our study shows that low dFe is the most likely cause of high chl a:heme b ratios in the IB, while a combination of low dFe and Si contributes to high chl a:heme b ratios in the SS. Furthermore, our results indicate that there are changes in the abundance of heme b proteins such as cytochromes in the phytoplankton communities in the IB and the SS as a result of low-nutrient concentrations. Changes in the abundance of cytochromes will affect electron transport and thus potentially influence the efficiency of photosynthesis and productivity of phytoplankton communities. Determination of particulate heme b in the ocean and comparison with POC, biomass, or chl a thus contributes to our understanding of the dynamics of phytoplankton blooms and adaptation of phytoplankton communities to prevailing growth conditions in the ocean.


[20] The authors gratefully acknowledge all the comments made by the anonymous reviewers of this work. This work was supported by a NERC Advanced Fellowship grant awarded to MG (NE/E013546/1). The authors would like to thank the officers, crew, and scientists of the R.R.S. Discovery and R.R.S. James Clark Ross during cruises D321, D326, and JR177. We would like to thank M. Patey for the nutrient data from the tropical North Atlantic; S. Henson and A. Poulton for the satellite images; A. Poulton for phytoplankton counts and taxanomic data; C.M. Moore and M. Lucas for particulate organic matter data; and C.M. Moore and T.S. Bibby for supplying the Fv/Fm data.