Biogeosciences Organic matter exudation by Emiliania huxleyi under simulated future ocean conditions

Emiliania huxleyi(strain B 92/11) was exposed to different nutrient supply, CO 2 and temperature conditions in phosphorus controlled chemostats to investigate effects on organic carbon exudation and partitioning between the pools of particulate organic carbon (POC) and dissolved organic carbon (DOC). 14C incubation measurements for primary production (PP) and extracellular release (ER) were performed. Chemical analysis included the amount and composition of high molecular weight ( >1 kDa) dissolved combined carbohydrates (HMW-dCCHO), particulate combined carbohydrates (pCCHO) and the carbon content of transparent exopolymer particles (TEP-C). Applied CO 2 and temperature conditions were 300, 550 and 900 μatm pCO2 at 14C, and additionally 900 μatmpCO2 at 18C simulating a greenhouse ocean scenario. Enhanced nutrient stress by reducing the dilution rate (D) from D = 0.3 d−1 to D = 0.1 d−1 (D =μ) induced the strongest response in E. huxleyi . At μ = 0.3 d−1, PP was significantly higher at elevated CO 2 and temperature and DO14C production correlated to PO 14C production in all treatments, resulting in similar percentages of extracellular release (PER; (DO 14C production/PP) × 100) averaging 3.74± 0.94 %. Atμ = 0.1 d−1, PO14C production decreased significantly, while exudation of DO14C increased. Thus, indicating a stronger partitioning from the particulate to the dissolved pool. Maximum PER of 16.3 ± 2.3 % were observed atμ = 0.1 d−1 at elevated CO2 and temperature. While cell densities remained constant within each treatment and throughout the experiment, concentrations of HMW-dCCHO, pCCHO and TEP were generally higher under enhanced nutrient stress. At μ = 0.3 d−1, pCCHO concentration increased significantly with elevated CO 2 and temperature. Atμ = 0.1 d−1, the contribution (mol % C) of HMWdCCHO to DOC was lower at elevated CO 2 and temperature while pCCHO and TEP concentrations were higher. This was most pronounced under greenhouse conditions. Our findings suggest a stronger transformation of primary produced DOC into POC by coagulation of exudates under nutrient limitation. Our results further imply that elevated CO 2 and temperature will increase exudation by E. huxleyiand may affect organic carbon partitioning in the ocean due to an enhanced transfer of HMW-dCCHO to TEP by aggregation processes.


Introduction
Primary production (PP) in the sunlit surface ocean is the driving force for the uptake of CO 2 and basis for its potential sequestration into the ocean's interior (Chisholm, 2000;Falkowski et al., 2000).The production of particulate organic carbon (POC) via PP is accompanied by the extracellular release (ER) of various amounts of dissolved organic carbon (DOC).ER represents a significant fraction of PP (Myklestad, 1977;Mague et al., 1980;Baines and Pace, 1991) and was reported from a variety of phytoplankton species in several studies over the last decades (Fogg, 1966;Myklestad, 1995;Biddanda and Benner, 1997;Descy et al., 2002;Wetz and Wheeler, 2007;López-Sandoval et al., 2011).Results from a modelling study with data from coastal, marine and estuarine systems revealed a linear relationship between PP and ER for non-freshwater systems with a percentage of extracellular release (PER) in a range of 2-50 % (Baines and Pace, 1991).A correlation between PP and ER was shown for irradiance (constant daylength) and temperature for marine Published by Copernicus Publications on behalf of the European Geosciences Union.
C. Borchard and A. Engel: Organic matter exudation by E. huxleyi (Verity et al., 1991) and freshwater algae (Zlotnik and Dubinsky, 1989) as long as applied conditions are in a range of optimum conditions of the tested algae.Data obtained under nutrient limitation frequently revealed PP and ER to be decoupled, generally leading to significantly higher PER (Myklestad, 1977;Obernosterer and Herndl, 1995;López-Sandoval et al., 2011).PER for marine phytoplankton typically ranges between 2-10 % (Nagata, 2000) or between 20-40 % in oligotrophic regions (Karl et al., 1998;Teira et al., 2001;López-Sandoval et al., 2011).This suggests that a fraction of photosynthesates (e.g., organic carbon) is not used for the build-up of biomass during nutrient limitation, but is discharged from the cell.Exudation of excess photosynthesates was suggested to be more energy-efficient for the phytoplankton cell than intracellular storage (Wangersky, 1978;Wood and van Valen, 1990).Maintaining the photosynthetic activity while bearing a loss of organic carbon may be additionally interpreted as a strategy to effectively shorten the lag phase for resuming the full photosynthetic capacity when nutrients become available again (Fogg, 1983).
Since preindustrial times, emissions of the greenhouse gas CO 2 into the atmosphere increased due to human activities (Caldeira and Wickett, 2003;Meehl et al., 2007).According to future projections on climate change, a rise in CO 2 and temperature will occur simultaneously (IPCC, Solomon et al., 2007) most likely accompanied by a reduced nutrient availability in the photic zone due to enhanced stratification of the ocean's surface (Boyd and Doney, 2002;Sarmiento et al., 2004).The sea surface is expected to be most affected by global change (Raven et al., 2005) and acidification and warming will likely influence the photosynthetic fixation of CO 2 and, therefore, the ocean's organic carbon pump and the fate of organic matter in the ocean (Falkowski, 1994;Arrigo, 2007;Caldeira, 2007).
Particulate and dissolved organic carbon production and therefore carbon fluxes were hypothesised to be affected by CO 2 and temperature (Engel, 2002;Leonardos and Geider, 2005;Moran et al., 2006;Wohlers et al., 2009;Kim et al., 2011).In accordance, TEP concentration has been shown to be related to seawater CO 2 -concentrations (Engel, 2002;Moran et al., 2006) and its formation was reported to be enhanced at elevated temperature (Claquin et al., 2008).Several studies report on enhanced exudation of TEP-precursors at nutrient stress (Staats et al., 2000;Passow, 2002a;Underwood et al., 2004), as a consequence of carbon overflow (Schartau et al., 2007).Factors that influence carbon overconsumption, exudation and TEP-formation will, thus, affect elemental ratios of particulate organic matter (POM) and may have strong implications for the organic carbon pump (Engel, 2002;Schneider et al., 2004).However, ER in referred studies was deduced from DOC accumulation rather than from direct rate measurements.
Emiliania huxleyi, a cosmopolitan species of coccolithophores, was the subject of several studies in the context of global change (Riebesell et al., 2000;Zondervan et al., 2002;Sciandra et al., 2003;Delille et al., 2005;Leonardos and Geider, 2005;Feng et al., 2008;Langer et al., 2009).Phytoplankton blooms dominated by Emiliania huxleyi were regularly observed at high irradiances in stratified waters where nitrate is abundant, but phosphate concentrations are well below 0.3 µmol l −1 (Veldhuis et al., 1994;Van der Wal et al., 1995;Head et al., 1998), in regions such as the North Sea to the east of the Shetland Islands or the Bjørnafjorden, Norway.E. huxleyi was frequently shown to grow well on low phosphate concentrations (Egge and Heimdal, 1994;Paasche, 2002) and was, therefore, suggested to be a strong competitor in phosphate depleted regions or situations in the ocean (Riegman et al., 2000).
E. huxleyi primary production was shown to be stimulated by elevated partial pressure of CO 2 (pCO 2 ) (Riebesell et al., 2000(Riebesell et al., , 2007;;Egge et al., 2009) and elevated temperature (Langer et al., 2007;Feng et al., 2008).Studies with E. huxleyi reporting on effects of nutrient limitation in the context of global change are scarce, but a clear decoupling of nutrient assimilation from carbon assimilation at high CO 2 conditions has been shown for nitrate (Sciandra et al., 2003) and phosphate limitation (Leonardos and Geider, 2005;Borchard et al., 2011).ER of organic carbon and the production of extracellular polysaccharides was reported for E. huxleyi previously (Nanninga et al., 1996;Biddanda and Benner, 1997;Godoi et al., 2009).Furthermore, the presence of coccolith polysaccharides (CP) (Westbroek et al., 1973), as well as the carbohydrate composition of CP in this species is well described (De Jong et al., 1979;Fichtinger Schepman et al., 1979;Kayano and Shiraiwa, 2009).Galacturonic acid, an acidic polysaccharide, was determined in E. huxleyi cultures (Nanninga et al., 1996) and is of great interest, since polysaccharides in general and acidic polysaccharides in particular, are known to facilitate TEP-formation (Mopper et al., 1995;Passow, 2000Passow, , 2002b)).
Results on the elemental composition of E. huxleyi in response to phosphorus limitation, CO 2 and temperature conditions during a chemostat experiment were presented in Borchard et al. (2011).High C : P ratios of particulate matter were observed during that study and were most pronounced at elevated CO 2 and combined high CO 2 and temperature.Elevated CO 2 also sustained higher biomass (e.g., cell densities) on identical nutrient supply.This led to a remarkable low P quota in high CO 2 cells potentially representing a minimum quota for sustaining growth at the given rate.
Here, we give detailed information on dissolved and particulate primary production derived from the same experiment in order to assess information on the partitioning of DOC and POC derived from E. huxleyi as a function of nutrient availability, CO 2 and temperature.The main objectives during the present study were to determine (i) the partitioning of photosynthetically produced POC and DOC, (ii) the amount and composition of high molecular weight (>1 kDa) dissolved combined carbohydrates (HMW-dCCHO) and particulate CCHO (pCCHO) and (iii) to test for linkages to TEPformation.

Experimental setup
Data shown here were obtained during the chemostat experiment described in more detail by Borchard et al. (2011).Briefly, a calcifying strain of Emiliania huxleyi (PML B92/11) was grown in CO 2 -and temperature-controlled chemostats with a culture volume of 9.2 l.Nutrient medium for the constant supply was prepared from sterile-filtered (Sartobran P, 0.2 µm capsule, Sartorius) natural seawater (SW) with a salinity of 32, pH of 7.97 and total alkalinity (TA) of 2281 µmol kg −1 SW.SW was enriched with nutrients to yield final concentrations of 29 µmol l −1 NO − 3 and 1.1 µmol l −1 PO 3− 4 and with a metal mix according to the f/2recipe (Guillard and Ryther, 1962).Medium for each chemostat was stored in 100 l reservoirs.In order to sterilize the medium, it was treated with UV irradiation (Microfloat 1/0, a.c.k.aqua concept GmbH) for 3 h before the addition of sterile-filtered f/2-vitamins (0.2 µm, Minisart, Sartorius).
After filling 9.2 l nutrient medium into each chemostat incubator, CO 2 and temperature conditions of 300 and 550 µatm at 14 • C and 900 µatm at 14 and 18 • C were applied and treatments are hereafter referred to as 300-14, 550-14, 900-14 and 900-18.CO 2 -concentrations in chemostats were achieved by constant aeration with gas obtained from a CO 2 regulation system based on the mixing of CO 2 free air and pure CO 2 using mass flow controllers (MFC, Type 1179 Mass Flo Controller; MKS Instruments, Germany) (Borchard et al., 2011).Briefly, premixed gases were channelled into each incubator via silicon and gas distribution tubes (ROBU; Type A, Por. 1) and allowed to equilibrate for 2 d before inoculation with cells.Each chemostat incubator was surrounded by a water jacket connected to a thermostat (Lauda, Ecoline Staredition, RE 104) setting temperatures to target values ±0.1 • C. A 16 h:8 h light:dark cycle with a photon flux density of 300 µmol photons m −2 s −1 was applied (TL-D Delux Pro, Philips; QSL 100, Biospherical Instruments, Inc.).
All pipes and tubes connected to the incubator were rinsed first with 10 % hydrochloric acid (HCl), subsequently with deionised water and then sterilized by autoclaving at 121 • C for 30 min before usage.The incubator and the reservoir vessels were cleaned with phosphate-free detergent, then soaked in 10 % HCl for 2 h and thoroughly rinsed with deionised water thereafter.
After the 2 days of CO 2 equilibration (considered as the first and second day of the experiment), cells were inoculated to each chemostat to a density of 3000 cells ml −1 and grown in batch mode for 3 d until the constant medium supply was applied for 12 d at a dilution rate (D) of D = 0.3 d −1 followed by 12 d at D = 0.1 d −1 .
Aeration with gas of preset CO 2 concentrations was maintained throughout the experiment in incubators and medium reservoirs.Gentle mixing at 50 rotations min −1 by a mechanical stirrer inside the incubator kept cells in suspension.Three samplings were accomplished during steady state of each growth rate at 8 a.m.(three hours after lights went on) and are referred to hereafter as sampling 1, 2 and 3 for µ = 0.3 d −1 (10, 14 and 17 days after the start of the experiment) and sampling 4, 5 and 6, for µ = 0.1 d −1 (22, 25 and 28 days after the start of the experiment).

Analytical methods
PO 14 C production (µmol l −1 d −1 ) was measured using the 14 C incubation method according to Steemann Nielsen (1952) andGargas (1975).50 ml samples were taken in triplicates from each chemostat, transferred into cell culture flasks (25 cm 2 , Corning) and spiked with approximately 5 µCi NaHCO − 3 (Hartmann Analytics, specific activity 40-60 mCi/mmole).Each triplicate set was incubated for 4 h at the same light and temperature conditions of the original chemostat treatment, but without aeration.Simultaneously, dark uptake was determined in triplicate for each treatment from 50 ml sample wrapped in aluminium foil and incubated in the dark for 4 h.Activity in the samples was determined by removing a 100 µl aliquot from three dark bottles prior to incubation and transferred to 6 ml liquid scintillation vials in which 200 µl of 2N NaOH were placed.4 ml of liquid scintillation cocktail were added thereafter.All incubations were stopped by gentle filtration on 0.4 µm polycarbonate filters (Nuclepore) at <150 mbar to avoid cell breakage.After rinsing with 10 ml sterile filtered seawater, filters were treated with fuming HCl in order to remove inorganic 14 C. Filters were transferred to 6 ml scintillation vials, 4 ml liquid scintillation cocktail (Ultima Gold AB) were added and samples were stored overnight before being counted in a Packard Tri Carb Liquid Scintillation Counter.Carbon incorporation rates were calculated in accordance to Gargas (1975) by implementing the values of total dissolved organic carbon determined on the respective sampling day for each chemostat.
Exuded dissolved organic carbon (DO 14 C production; µmol l −1 d −1 ) was measured in the 0.4 µm-filtrate obtained during the PO 14 C-filtration.4 ml liquid sample were acidified with 100 µl 2 M HCl and left for 7 days in a desiccator with a NaOH-trap at ∼600 mbar depression to allow for outgassing of inorganic 14 C.After the addition of 15 ml liquid scintillation cocktail (Ultima Gold AB), overnight storage, counting and calculations were accomplished as described above.
Primary Production (PP; µmol C l −1 d −1 ) was derived from the sum of the production of PO 14 C and DO 14 C (PO 14 C production + DO 14 C production = PP).The percentage of extracellular release (PER) was calculated as (DO 14 C production/PP) × 100).
Transparent exopolymer particles (TEP) were determined according to the colorimetric method by Passow and Alldredge (1995).10 to 20 ml sample were filtered in triplicate onto 0.4 µm polycarbonate filters (Nuclepore), stained with Alcian Blue and rinsed with several ml of ultrapure water.Filters were stored in polypropylene tubes at −20 • C until analysis.TEP-concentrations are given in units of µg Xanthan-equivalents per litre (µg X eq.l −1 ).Conversation to carbon-units (TEP-C) in µmol C l −1 was done by applying a factor of f 0 = 0.63 as suggested by Engel et al. (2004a).Alcian Blue was shown to stain the surface of E. huxleyi coccoliths.Therefore, TEP-C concentrations were corrected for the amount of Alcian Blue per cell as described by Engel et al. (2004a).
Cell densities were determined daily as the mean of three consecutive measurements of 500 µl sample with a Beckman Coulter Counter (Coulter Multisizer III) equipped with a 100 µm aperture.Prior to measurements, samples were diluted 1:100 with 0.2 µm pre-filtered (Minisart, Sartorius) seawater with a salinity of 32.The equivalent spherical diameter (ESD) of E. huxleyi was determined to be within a particle size range of 3.27-7.96µm, based on microscopic measurements.A constant cell density over three consecutive days (maximum variation of 10 %) was accepted as criterion for steady state growth in continuous cultures after Leonardos and Geider (2005).
During the present study, bacterial cell numbers were checked by flow cytometry and numbers were low and similar in all treatments.We estimated the carbon demand from growth rate and an assumed biomass of 20 fg C bacterial cell −1 (Lee and Fuhrman, 1987), a growth efficiency of 30 % and assuming that 70 % of carbon uptake was respired.Relative to the huge concentrations of organic carbon produced by E. huxleyi, the bacterial carbon demand was negligible for the present study.
The carbonate chemistry during the experiment is described in more detail in Borchard et al. (2011) and was determined from pH and total alkalinity (TA).Briefly, pH values in the treatments were measured with a proton sensitive combined pH-temperature electrode (Sentix 41; WTW standard DIN/NBS buffers PL 4, PL 7 and PL 9) and recorded hourly on a data logger (WTW; pH 340i) throughout the experiment.For TA, 200 ml of sample were filtered through combusted (8 h, 500 • C) glass fibre filters (GF/F, Whatman) and analysed by the titration method with 0.1 M HCl (Gran, 1952).pCO 2 , total dissolved inorganic carbon (DIC), carbonate (CO 2− 3 ), bicarbonate (HCO − 3 ) and calcite in each treatment and medium reservoirs were calculated from pH (NBS) and TA values at the beginning of the experiment and at each sampling day using the programme CO2sys (Lewis and Wallace, 1998).After equilibration with distinct CO 2 -gas concentrations, initial pH values ranged from 8.29 to 7.86 at a TA value of 2281 µmol kg −1 seawater, representing a CO 2 range from 287 ± 6 and 897 ± 17 µatm CO 2 .Carbonate chemistry was strongly affected by biological activity of E. huxleyi, but remained constant within each dilution rate (Table 1).The complete dataset for the carbonate chemistry and related discussion is given elsewhere (Borchard et al., 2011, in Sect. 4.3 and Tables 1 and 3).

Data treatment
During steady state growth, the total production rates (γ [µmol l −1 d −1 ]) for CCHO and TEP-C were calculated from concentrations (x [µmol l −1 ]) and growth rates (µ [d −1 ]) by applying the following equation: Cell normalised production rates (γ cell ; fmol cell −1 d −1 ) were derived from cell normalised concentrations of CCHO and TEP-C (x cell ; fmol cell −1 ) with n being the cell density by the use of the following equation: Mean values ± standard deviations for each dilution rate were derived from sampling 1, 2 and 3 (µ = 0.3 d −1 ) and 4, 5 and 6 (µ = 0.1 d −1 ), respectively.Contributions (mol %) and elemental ratios are consistently derived from concentrations (mol l −1 : mol l −1 ).Differences between treatments were tested by means of analysis of variance (two-way ANOVA), and by means of t-tests.Statistical treatment of data was performed by using the software package SigmaPlot 10.01 (Sys-Stat).

Nutrients and cell densities
Supply media of all chemostats was enriched to identical nutrient concentrations of 29 µmol l −1 NO − 3 and 1.1 µmol l −1 PO 3− 4 .Data on cell densities and elemental stoichiometry in the course of the experiment are given in more detail in Borchard et al. (2011).Briefly, growth of E. huxleyi was stimulated by a rise in CO 2 and temperature during the initial nutrient replete batch phase and during the transition to phosphorus limitation.While cell densities within each treatment remained constant throughout the experiment after reaching the steady state (5 to 7 days after setting the dilution rate (D) to D = 0.3 d −1 ), cell densities differed significantly (ttest, n = 17, p < 0.001) between treatments ranging from 3138×10 5 ±8.1 % cells l −1 in the 300-14 treatment and maximum 4238 × 10 5 ± 7.5 % cells l −1 in the 900-18 treatment (Table 1).Cell densities remained stable throughout both dilution rates with a maximum variation of 10.2 % (550-14), indicating that growth rates balanced dilution rates with µ = D = 0.3 d −1 and µ = D = 0.1 d −1 .PO 3− 4 and dissolved organic phosphorus concentrations were below the detection limit in all treatments for both growth rates.Residual DIN concentrations of 0.42-1.08µmol l −1 were determined in culture media of all treatments, indicating exclusive phosphorus depletion.Average PON:POP ratios were 25.5 ± 3.0 and similar for all treatments.Cell quota for POP (P quota) exhibited a decreasing trend with increasing CO 2 and temperature reflecting the differences in cell densities produced on identical nutrient concentrations (Table 1).A minimum P quota of 2.19±0.08fmol cell −1 was determined for cells growing at D = 0.1 d −1 (900-18).

Carbohydrate composition
Carbohydrate composition was affected by changes in nutrient supply and by varied CO 2 and temperature conditions, albeit to a lesser extent (Fig. 3).
At each growth rate, TEP-C was significantly higher in 900-18 than in 300-14 and 550-14 (p < 0.002).In accordance with total concentrations of TEP-C (µmol l −1 ), cell normalised TEP-C (fmol cell −1 ) was significantly higher at µ = 0.1 d −1 compared to µ = 0.3 d −1 in all treatments (two way ANOVA, Holm-Sidak method, p < 0.001).Cell normalised TEP-C was highest in 900-18 at both growth rates, although not statistically different from values derived from the other treatments (p = 0.314) (Table 2).Throughout the experiment, pCCHO significantly correlated with TEP-C (Pearson Product Moment Correlation; n = 24, r 2 = 0.826, p < 0.001) (Fig. 5).Furthermore, concentrations of DO 14 C (µmol C l −1 ) also significantly correlated to TEP-C (Pearson Product Moment Correlation; n = 24, r 2 = 0.862, p < 0.001) (Fig. 6).This strongly suggests coherence between the presence of dissolved precursors and TEP.The fact, that higher exudation is not mirrored in higher HMW-dCCHO is most likely reasoned by a transformation of dissolved components into the particulate pool in form of TEP, rich in pC-CHO.

Particulate and dissolved primary production
In the present experiment, while growth of E. huxleyi adapted to the nutrient supply by lowering the growth rate, primary production (PP) and, therefore, the basis for carbon metabolism was clearly decoupled from nutrient assimilation.PP partitioning in dissolved and particulate organic carbon is highly variable, and percentages of extracellular release (PER), depend strongly on environmental conditions that affect the cell's physiology (Myklestad and Haug, 1972;Zlotnik and Dubinsky, 1989;Chen and Wangersky, 1996;Penna et al., 1999;Staats et al., 2000;Wetz and Wheeler, 2003;Magaletti et al., 2004).
During this study, PP of E. huxleyi was highest in the 900-18 treatment at µ = 0.3 d −1 (Fig. 1, upper panel), showing a stimulation of PP by the combined rise of CO 2 and temperature.This is in accordance with earlier findings on the effect of elevated CO 2 (Riebesell et al., 2000) or temperature (Langer et al., 2007) on E. huxleyi obtained under nutrient   ally assumed (Zlotnik and Dubinsky, 1989;Baines and Pace, 1991;Verity et al., 1991).Nutrient limitation, however, was regularly shown to decouple PP and ER, resulting in higher PER due to a strong relative increase in ER (Myklestad, 1977;Obernosterer and Herndl, 1995;Staats et al., 2000;López-Sandoval et al., 2011).While high light intensities and CO 2 stimulate carbon assimilation (e.g., primary production) nutrient limitation hampers cell division and growth at the same time (Fogg, 1966;Baines and Pace, 1991;Banse, 1994).A particular increase in carbon assimilation relative to the uptake of N and P is referred to as carbon overconsumption (Toggweiler, 1993;Schartau et al., 2007).Carbon overconsumption susceptibly alters the stoichiometric relation between major elements and large deviations from Redfield's C:N:P of 106:16:1 were regularly observed in the field (Fraga, 2001;Geider and LaRoche, 2002) as well as in laboratory experiments (Sciandra et al., 2003;Leonardos and Geider, 2005).
The present experiment was designed to investigate the combined effect of nutrient limitation (e.g., phosphorus limitation), CO 2 and temperature on organic carbon production at two low growth rates.In agreement with earlier findings, higher PER was determined at conditions of enhanced nutrient stress after changing the nutrient supply rate from D = 0.3 d −1 to D = 0.1 d −1 .A direct relationship between PER and growth rate was questioned (Williams, 1990), primarily following a chemostat study with Thalassiosira pseudonana (diatom) (Smith and Platt, 1984) exhibiting very low PER varying in a range of 1.2-2.7 % at growth rates between 0.29 and 2.50 d −1 .Data obtained here, however, showed a strong impact of growth rate as a result of enhanced nutrient limitation on PER, which was obvious in the pronounced increase of PER from averaged 3.7 % at µ = 0.3 d −1 to 11.0 % at µ = 0.1 d −1 (Fig. 1, lower panel).Since the reduction in growth rate is a result of enhanced nutrient limitation, data presented here are in accordance with a variety of studies reporting on higher PER with increasing nutrient limitation (Myklestad, 1977;Obernosterer and Herndl, 1995;López-Sandoval et al., 2011).
The effect of elevated CO 2 and temperature on phytoplankton exudation, however, has not yet been directly tested, but ER of organic molecules was hypothesised to be stimulated by either elevated CO 2 (Engel, 2002) or by elevated temperature (Moran et al., 2006;Wohlers et al., 2009;Engel et al., 2011).In the present study, both, PP and ER showed an increasing trend with elevated CO 2 and temperature at a growth rate of µ = 0.3 d −1 (Fig. 1), and the coupling of PP and ER, thus, yielded comparable PER in all treatments.At enhanced nutrient stress (µ = 0.1 d −1 ) the CO 2 and temperature effect on POC production was diminished (Fig. 1, upper panel).DOC production, however, was clearly stimulated by the combined effect of enhanced nutrient limitation and elevated CO 2 and temperature (Fig. 1, middle panel).Organic carbon production was, thus, shifted to the dissolved pool at greenhouse conditions, yielding highest PER of 16.3 % (Fig. 1, lower panel).At severe nutrient stress, ER is, therefore, most likely of greater importance for the carbon partitioning because the reduction of POC production coincides with an enhancement of DOC production.For the future ocean, this implies a potential decoupling of growth, POC production and DOC production.
In natural systems, freshly produced DOC is readily taken up by heterotrophic bacteria (Azam et al., 1983;Martin et al., 1987).Elevated temperature as well as decreased pH has been shown to accelerate enzymatic reactions and to enhance bacterial degradation in marine systems especially concerning protease, and α-and β-glucosidase activity (Grossart et al., 2006;Piontek et al., 2009Piontek et al., , 2010)).Yamada and Suzumura (2010) reported that bacterial β-glucosidase activity is unaffected by ocean acidification and amino peptidase activity even decreased at lowered pH.A recent publication by Endres et al. (2012) however, supported the hypothesis on stimulated bacterial enzyme activity under future ocean conditions by reporting on an increased amino peptidase activity in autotrophic cyanobacteria.Phytoplankton exudates derived under phosphorus limitation, however, may in general not be efficiently utilised by heterotrophic bacteria.This was either argued by a severe slowing down of bacterial metabolism due to limiting inorganic and organic phosphorus concentrations (Obernosterer and Herndl, 1995), or by the recalcitrant characteristics of DOC released from phytoplankton under phosphorus limitation (Puddu et al., 2003).A loss of DOC due to heterotrophic activity might, thus, be minimised and carbon overconsumption and excess exudation may result in higher concentrations of DOC in the future ocean.

Carbohydrates and linkages to TEP-formation
Concentrations of HMW-dCCHO, pCCHO and TEP-C were higher in all treatments under enhanced nutrient limitation (µ = 0.1 d −1 ).Specific production rates of CCHO reached physiological steady states for each D, albeit, decoupled from cell growth, resulting in different yields for µ = 0.3 d −1 and µ = 0.1 d −1 , respectively.Relative to the growth rate, specific production of carbohydrate-rich compounds was generally higher at µ = 0.1 d −1 , resulting in higher concentrations of all carbon rich compounds.The different yields of organic carbon in form of carbohydrates and TEP-C were thus not related to the cell density of E. huxleyi, but clearly due to changed specific production (Figs. 2 and 4).
A stronger partitioning of PP to the dissolved pool may reduce carbon export in the future ocean if DOC is not transferred to the particulate pool via aggregation processes (Kim et al., 2011).Phytoplankton exudation, however, does not only increase DOC concentration, it also provides the precursors for the formation of transparent exopolymer particles (TEP) by physical coagulation (Passow, 2000).This abiotic formation of particles such as TEP from organic molecules was proposed to be a major mechanism for the loss of dissolved carbohydrates in the ocean (Skoog and Benner, 1997;Engel et al., 2004b).A recent publication by Passow (2012) showed that the abiotic TEP formation, i.e., the coagulation process itself, is not impacted by changes in pH.Here we report on the biological production of dissolved precursors and the subsequent production of TEP that was significantly related to PER (Fig. 6).The composition of dissolved carbohydrates was significantly affected by varied conditions of nutrient supply while the influence of CO 2 and T was negligible (Fig. 3).PER in all treatments was higher at µ = 0.1 d −1 compared to µ = 0.3 d −1 and highest exudation was clearly determined for the cells grown at elevated CO 2 and temperature (Fig. 1).The higher TEP formation observed at 900-18 is, therefore, a likely result of higher concentrations of precursors due to enhanced biological activity, rather than the quality of dissolved material.For the present study, the steep rise in PER, as a consequence of changed nutrient supply, CO 2 and temperature, was not mirrored in the proportion of HMW-dCCHO contributing to DO 14 C (Fig. 7, lower panel) and of HMW-dCCHO to tCCHO (Fig. 2, lower panel).The TEP-C:PO 14 C ratio, however, (Fig. 7, upper panel) resembles the PER.A significant correlation between the concentration of dissolved precursors (DO 14 C, µmol C l −1 ), and TEP-C (µmol C l −1 ) was determined (r 2 = 0.862, p < 0.001, Fig. 6).The transfer of HMW-dCCHO (low HMW-dCCHO:DO 14 C) and the concomitant relative increase in pCCHO (low HMW-dCCHO:pCCHO) was accompanied by an increase in TEP-C (high TEP-C:PO 14 C) (Figs. 4 and 7).These finding, therefore, indicate a fast transformation of HMW-dCCHO to pCCHO in form of TEP, which was most pronounced at elevated CO 2 and temperature at enhanced nutrient stress.
Results obtained during this study, thus, strongly support the idea of a direct pathway between DOC in form of HMW-dCCHO and POC by physical aggregation, as suggested previously (Chin et al., 1998;Wells, 1998;Engel et al., 2004b;Verdugo et al., 2004;Verdugo et al., 2008;Ding et al., 2009) and shows, that this pathway will be of greater importance under conditions of simultaneously increased CO 2 and temperature.The hypothesis of a direct transfer of HMW-dCCHO to TEP is additionally supported by a significant correlation between pCCHO and TEP-C concentration (r 2 = 0.826, p < 0.001) (Fig. 5).Compared to µ = 0.3 d −1 , HMW-dCCHO composition at µ = 0.1 d −1 was shifted to comparable contributions of Man/Xyl, Ara/GalN and Glc accompanied by high contributions of Glc, Man/Xyl and Gal-URA to pCCHO and an increase in TEP (Figs. 3 and 4 and Table 3).Compared to other carbohydrate monomers, this indicates a higher relative increase of the production and exudation of Man/Xyl and Gal-URA at excess nutrient limitation.Throughout the experiment for all treatments, concentrations of particulate Gal-URA, an acidic sugar which is typical for E. huxleyi (Nanninga et al., 1996), correlated significantly with concentrations of TEP-C (r 2 = 0.827, p < 0.001).Low contributions of HMW-dCCHO to pCCHO despite the high ER, can be explained by a transfer from neutral (Man/ Xyl) and acidic HMW-dCCHO (Gal-URA) to higher proportions of neutral and acidic pCCHO as a characteristic fraction of TEP (Skoog et al., 2008).Therefore, data obtained here show coherence between enhanced exudation of DOC and higher concentrations of pCCHO and TEP-C.A significant proportion of DOC (up to 90 %) derived from phytoplankton exudation was shown to be comprised by carbohydrates (Myklestad, 1995).Here, HMW-dCCHO comprised 12 to 43 % of freshly produced DOC (Fig. 7) and the comparably low proportions further indicate the transfer of HMW-dCCHO to POC due to aggregation processes in accordance to the model proposed by Engel et al. (2004b) (Figs. 2, 6 and 7).
Earlier studies with E. huxleyi reported that TEP-C accounted for 63 % of POC (Engel et al., 2004a) during a mesocosm experiment.In the Bay of Biscay, Harlay et al. (2009) observed a 12 % contribution of TEP-C to POC during a bloom of coccolithophores, dominated by E. huxleyi.During this study, TEP-C explained on average 3 % at µ = 0.3 d −1 and 8 % at µ = 0.1 d −1 of POC with highest proportions of TEP-C contributing to POC observed in the 900-18 treatment.
Thus, compared to field and mesocosm systems, the contribution of TEP-C to POC was relatively low during this chemostat experiment.This suggests that physiological nonsteady state situations, such as occurring naturally during bloom situations, have a different impact than conditions of steady-state growth.Clearly, more information on loss processes of TEP in marine systems is needed to evaluate the role of TEP for carbon partitioning in the future ocean.

Conclusions
Emiliania huxleyi is one of the most abundant coccolithophore in the global ocean and despite its small size an important contributor to marine primary production.Significantly higher cell densities and, therefore, minimised cell quota for phosphorus were exclusively induced by elevated CO 2 (Borchard et al., 2011).Sustaining photosynthesis and growth, while increasing DOC production, potentially represents a physiological strategy allowing (1) the production of high biomass at high CO 2 on low nutrient supply and (2) a quick response of cellular processes when nutrients become available again.
Data presented here reveal that nutrient concentration had the strongest impact on organic carbon partitioning, relative to the influence of CO 2 and temperature.DOC and POC production could, therefore, be more affected in the future ocean by changed nutrient or growth conditions, either on a regional basis or during the seasonal succession.Our data further imply that partitioning of organic carbon production to the dissolved pool will most likely be additionally promoted at elevated CO 2 and temperature.This, however, will depend on the synergistic effect of nutrient limitation and elevated CO 2 and temperature on P quota, cell size and carbon overconsumption.
Despite the fact, that DOC itself does not sink, an enhanced DOC pool most likely provides higher concentrations of precursor material in form of HMW-dCCHO for aggregation processes.Although only slightly affected by elevated CO 2 , the composition of HMW-dCCHO could be relevant for an efficient transfer of HMW-dCCHO to the particulate pool.However, TEP-formation in the greenhouse ocean seems to be primarily driven by the higher concentration of precursor material due to higher exudation.
The importance of TEP as part of the POC pool is well acknowledged in studies of organic carbon fluxes (Engel and Passow, 2001;Mari et al., 2001;Engel et al., 2004b;Arrigo, 2007;De La Rocha and Passow, 2007;Schartau et al., 2007;Verdugo et al., 2008).A quantitative and qualitative connection between DOC production and TEP-C formation as shown for the present study was proposed earlier (Jackson and Burd, 1998;Engel et al., 2004b).However, timing of the biological process of extracellular release and the abiotic formation of particles may potentially be decoupled.Low HMW-dCCHO concentrations contributing to tC-CHO at high exudation rates may, thus, point to a fast aggregation of DOC into TEP.Since precursors in form of dissolved CCHO are effectively reduced at elevated CO 2 and temperature, data obtained from this experimental study suggest, that TEP-formation occurs faster at elevated temperature most likely due to physical and chemical processes (Mc-Cave, 1984;Claquin et al., 2008).However, the quantitative importance of TEP contributing to the POC pool in the future ocean needs further studies taking into account the multiple factors and processes occurring in natural ecosystems.

3423, 2012 C. Borchard and A. Engel: Organic matter exudation by E. huxleyi shifted
to the particulate pool with increasing CO 2 and temperature.