A Universal Carbonate Ion Printer-friendly Version Interactive Discussion a Universal Carbonate Ion Effect on Stable Oxygen Isotope Ratios in Unicellular Planktonic Calcifying Organisms a Universal Carbonate Ion Printer-friendly Version Interactive Discussion

Biogeosciences Discussions This discussion paper is/has been under review for the journal Biogeosciences (BG). Please refer to the corresponding final paper in BG if available. Abstract The oxygen isotopic composition (δ 18 O) of calcium carbonate of planktonic calcifying organisms is a key tool for reconstructing both past seawater temperature and salinity. The calibration of paloeceanographic proxies relies in general on empirical relationships derived from experiments on extant species. Laboratory experiments have more 5 often than not revealed that variables other than the target parameter influence the proxy signal, which makes proxy calibration a challenging task. Understanding these secondary or " vital " effects is crucial for increasing proxy accuracy and possibly for developing new biomarkers. We present data from laboratory experiments showing that oxygen isotope fractionation during calcification in the coccolithophore Calcidiscus 10 leptoporus and the calcareous dinoflagellate Thoracosphaera heimii is dependent on carbonate chemistry of seawater in addition to its dependence on temperature. A similar result has previously been reported for planktonic foraminifera, suggesting that the [CO


Introduction
Calcification by marine organisms is a key component of the global carbon cycle.Planktonic calcifying organisms are particularly significant since they are responsible for more than 80 % of global CaCO 3 production and a high proportion of their shells is exported to the seafloor to form carbonate sediments (Milliman, 1993).
The unicellular coccolithophore algae, members of the division Haptophyta, are among the major primary producers and calcifiers worldwide.Coccolithophores are present in virtually all photic zone marine environments (Baumann et al., 2005;Thierstein and Young, 2004) and together with planktic foraminifera they dominate carbonate production in the open ocean (Ziveri et al., 2007;Broecker and Clark, 2008).Another group of planktonic calcifying organisms is a monophyletic lineage of peridiniphycidean dinoflagellates that live in the upper water column where light is available for photosynthesis and during their life cycle produce cysts that are characterized by the incorporation of calcite in at least one layer of the cyst wall.
The chemical composition of the calcite produced by these organisms (i.e., foraminiferal tests, the coccoliths of coccolithophores and the cysts of calcareous dinoflagellates) provides a unique source of information with regard to paleoclimatological studies (e.g., Erez and Luz, 1983;Stoll and Ziveri, 2004;Katz et al., 2010) The oxygen isotope composition of carbonate fossils has been used to reconstruct the most sought-after target in paleoceanography, the sea surface temperature (SST).The δ 18 Otemperature relationship is well characterized in foraminifera (Bemis et al., 1998) and has also been studied in coccolithophores (Ziveri et al., 2003 and references therein) and calcareous dinoflagellate cysts (Zonneveld et al., 2007).In addition to the temperature effect, a dependence of oxygen isotope fractionation on the carbonate chemistry of seawater was demonstrated in four planktic foraminiferal species (Spero et al., 1997;Bijma et al., 1999).A decrease in fractionation with increasing carbonate ion concentration was recorded in tests of these species, but the slopes for this relationship differed by a factor of up to two.The more gentle of the slopes was explained in terms of the isotopic composition of the sum of the carbon species in seawater (Zeebe, 1999), but the steeper slope has remained unexplained to date.
The influence of carbonate chemistry on the oxygen isotope fractionation of coccolithophores and calcareous cyst producing dinoflagellates has never been studied.We therefore conducted controlled laboratory experiments using one of the most important calcite producers among the coccolithophores, Calcidiscus leptoporus (Ziveri et al., 2007), as well as the most abundant living calcareous dinoflagellate, Thoracosphaera heimii.The carbonate chemistry of seawater was altered to determine its impact on oxygen isotope fractionation during calcification.Based on our experimental data, we propose a simple conceptual model that can explain the different δ 18 O slopes in terms of a different contribution of HCO − 3 conversion to CO 2− 3 to the establishment of supersaturation with respect to calcite in the calcification vesicle.This model is subsequently applied to explain the data previously obtained from studies on planktic foraminifera (Spero et al., 1997;Bijma et al., 1999).

Experimental design and sampling
Clonal cultures of Calcidiscus leptoporus (strains RCC1154, formerly known as AC360 and AS31 from the Alboran Sea off Spain, and RCC1135 formerly known as AC365 and NS6-1 from the South Atlantic off South Africa) and of T. heimii (strain RCC1511 formerly known as JF1 from the Pacific Ocean) from the Roscoff Culture Collection, (http: //www.sb-roscoff.fr/Phyto/RCC/)were grown in sterile filtered (0.2 µm) seawater enriched with 100 µM nitrate and 6.25 µM phosphate and with trace metal and vitamin sup-plements according to f/2(-Si) (Guillard and Ryther, 1962).The incident photon flux density was 350 µmol m −2 s −1 and a 16/8 h light/dark cycle was applied.Experiments were carried out at 20 • C for C. leptoporus and 17 • C for T. heimii.Cells were acclimated to experimental conditions for approximately 10 generations and grown in dilute batch cultures in triplicate (Langer et al., 2006).Low cell density at harvest (in general less than 6000 cells per ml) resulted in less than 8 % DIC (dissolved inorganic carbon) consumption (i.e., DIC consumed by the cells at the end of experiment) and a shift in pH of no more than 0.06 units.
The carbonate system of seawater can be manipulated in various ways.Changes in atmospheric pCO 2 can be used to alter [CO 2 ], pH and DIC, with TA (total alkalinity) remaining constant.Manipulating pCO 2 requires bubbling with CO 2 gas (constantly throughout the experiment if an open system is employed).Alternatively, the addition of the acid HCl or the base NaOH can be used to adjust [CO 2 ], pH, and TA, with DIC remaining constant.Both methods, TA manipulation (C.leptoporus) and DIC manipulation (T.heimii), were employed in this study, in both cases in closed systems (see below).It should be noted that chemical changes imposed by the two methods are of similar magnitude.The pCO 2 range typically used in experiments is ∼180 to ∼1000 µatm.This range can be covered by either increasing DIC by approximately 15 % or decreasing TA by approximately 15 % (to increase [CO 2 ]).The change in either DIC or TA is small compared to the changes in CO 2 concentration at the upper limit (increase by a factor of ∼6), CO 2− 3 concentration (decrease by a factor of ∼4), and H + concentration (increase by a factor of ∼5, i.e., a pH drop of ∼0.7 units) (for details on the carbonate system refer to Zeebe and Wolf-Gladrow, 2001).Experiments in which DIC is altered should thus be comparable to experiments in which TA is altered (Schulz et al., 2009;Hoppe et al., 2011).
In order to prevent gas exchange with the atmosphere in our experiments, 2.4 L flasks were filled without headspace and closed with Teflon-lined screw caps.Determination of cell density, however, required regular sampling for cell counts, thereby creating a maximum headspace of 6 ml, and resulting in a negligible shift (3 %) in CO 2 (aq) concentration.Samples for alkalinity measurements were filtered (0.6 µm pore size), poisoned with 1 mL of a HgCl 2 solution (35 g l −1 ) and stored in 300 mL borosilicate flasks at 0 • C. DIC samples were sterile filtered using a syringe filter (0.2 µm) and stored in 13 mL borosilicate flasks free of air bubbles at 0 • C. Total alkalinity was calculated from linear Gran plots (Gran, 1952) after duplicate potentiometric titration (Bradshaw et al., 1981;Brewer et al., 1986) and DIC was measured photometrically (Stoll et al., 2001) in triplicate.Precision of the total alkalinity measurements was ∼3 µmol L −1 and accuracy ∼4 µmol L −1 .For DIC, precision was ∼4 µmol L −1 and accuracy ∼5 µmol L −1 .The carbonate system was calculated from temperature, salinity, and the concentrations of DIC, total alkalinity and phosphate, using the CO 2 SYS software (Lewis and Wallace, 1998).Equilibrium constants of Mehrbach et al. (1973) refitted by Dickson and Millero (1987) were chosen.
Cell samples for oxygen and carbon isotope analysis were centrifuged in 50 mL Falcon tubes, centrifuged again in 1 mL tubes in order to remove seawater, dried at 60 • C for 48 h, and stored at room temperature.After the last centrifugation step, care was taken to remove as much of the seawater medium as possible in order to minimize contamination due to residual salts.Seawater samples for isotope analysis were sterile filtered (0.2 µm), poisoned with HgCl, and stored in gas tight bottles at room temperature prior to analysis.For determination of cell density, samples were taken daily or every other day, stored at 0 • C and counted within 3 h of sampling using a Sedgwick Rafter counting cell.

Sample preparation and measurement of isotope ratios
The δ 18 O of water samples from the C. leptoporus experiments was measured at the Leibniz Laboratory, Kiel University, on a Finnigan Delta E connected with an equilibration bath (Equi).An aliquot of the sample water was isotopically equilibrated with an external CO 2 , this CO 2 was extracted, analyzed by mass spectrometer, and δ 18 O calculated from the isotope composition.δ 18 O calibration of the working reference gas was based on VSMOW, with a fine correction derived from analysis of VSMOW, GISP, and SLAP.This calibration was performed once or twice a year.The stability control in routine operation used internal water standards of different isotopic composition.As with the DIC sample series, a second gas standard was measured at the end of each sample batch.The 1σ standard deviation of water samples was ±0.05 for δ 18 O.The poisoned filtered medium samples from the T. heimii culture experiments were measured at the Vrije Universiteit (Amsterdam) using an on-line gas preparation (Finnigan GasBench II).The 1σ standard deviation of water samples for δ 18 O was ±0.05.Reproducibility tests were based on a routinely run laboratory water standard (Vienna-Standard Mean Ocean Water (V-SMOW)).
The stable isotopes of coccolith and calcareous dinoflagellate samples were measured at the Vrije Universiteit (Amsterdam) using a CARBO-KIEL automated carbonate preparation device linked on-line to a Finnigan MAT252 mass spectrometer (Kiel device).The reproducibility of a routinely analyzed carbonate standard (GICS) was better than 0.15 ‰ for δ 18 O for the MAT252.Based on replicate analyses of splits of culture samples, the mean reproducibility of the δ 18 O measurements was better than ±0.2 ‰.

Experimental results and discussion
Carbonate ion concentration in the seawater medium in which the cells were grown, calculated using CO 2 SYS (Lewis and Wallace, 1998), ranged from 78 to 530 µmol kg −1 for C. leptoporus experiments and from 117 to 239 µmol kg −1 for T. heimii experiments.The δ 18 O in coccoliths of C. leptoporus was inversely correlated with the computed [CO 2− 3 ] with a regression slope of −0.0048 (‰ (µmol kg −1 )) (r 2 = 0.89), (Fig. 1).The δ 18 O in T. heimii calcite exhibited the same type of inverse relationship with [CO 2− 3 ], but with a regression slope of −0.0243 (‰ (µmol kg −1 )) (r 2 = 0.95).The 1σ confidence bounds are shown in Fig. 1.Oxygen isotope fractionation in these two phylogenetically distant unicellular planktonic calcifying taxa, as well as in two planktonic Foraminifera (Spero et al., 1997;Bijma et al., 1999), thus consistently decreases with increasing [CO 2− 3 ] of seawater.This relationship suggests a common oxygen isotope fractionation mechanism.We propose below a conceptual model that can explain the different slopes identified in this and previous studies based on the contribution of HCO − 3 to the CO 2− 3 pool in the calcifying vesicle.
In coccolithophores, precipitation of calcite takes place intracellularly within a calcifying vesicle, a membranedelimited space that is completely isolated from the cytoplasm (e.g., Young et al., 1999).While the pH in the external medium can change under different environmental conditions, cellular pH-homeostasis should keep the pH inside the vesicle constant at an alkaline value favorable for calcite precipitation.
Interestingly, a similar vesicle-based calcification mechanism has been proposed for the common calcareous dinoflagellate T. heimii (Inouye and Pienaar, 1983).Although planktic foraminifera are thought to calcify in an extracellular space (Spero, 1988), the role of seawater endocytosis in the biomineralization process in calcareous foraminifera (Bentov et al., 2009) (Spero et al., 1997), and G. bulloides (Spero et al., 1997) (lines are linear regressions fitted to the data).The regression slopes are −0.0022‰ (µmol kg −1 ) −1 for O. universa (HL = high light; DC = dark conditions), −0.0048 ‰ (µmol kg −1 ) −1 for C. leptoporus, −0.0047 ‰ (µmol kg −1 ) −1 for G. bulloides, and −0.024 ‰ (µmol kg −1 ) −1 for T. heimii.1σ confidence bounds are shown for each regression (shaded areas).results in T. heimii presented in Gussone et al. (2010) showing strong resemblance with values recorded in planktonic foraminifera thus, supporting the similarities in calcification mechanisms proposed by our model.The extracellular calcification space of foraminifera is isolated from the seawater by a so called pseudopodial network, so that, in effect, also foraminifera calcify in a space which is isolated by means of plasmamembrane not only from the seawater, but also from the cytoplasm (Erez, 2003).This common basic feature of the calcification of the three phylogenetically distinct groups of calcifiers, coccolithophores, foraminifera, and dinoflagellates, can partly account for the fact that it is possible to formulate one single model explaining the dependency of δ 18 O on carbonate chemistry as will be discussed below.Zuddas and Mucci (1998) and Kim et al. (2006) demonstrated that for an alkaline pH, the kinetics of CaCO 3 precipitation are dominated by the reaction: Ca 2+ + CO 2− 3 = CaCO 3 .Assuming that calcium carbonate is precipitated mainly from carbonate, the δ 18 O composition of the precipitate simply reflects the isotopic composition of CO 2− 3 in the vesicle.Since at alkaline pH values CO 2− 3 is the major carbon source for CaCO 3 it is seems reasonable to assume that the uptake of both CO 2− 3 and HCO − 3 from the surrounding medium contributes to the CO 2− 3 pool in the vesicle.The uptake of CO 2− 3 is assumed to yield a pool size proportional to [CO 2− 3 ] in the external medium (ext): f ×[CO 2− 3 ] ext with f ≤ 1. Inorganic carbon transported into the vesicle (V) as HCO − 3 is converted to CO 2− 3 , and the [CO 2− 3 ] V increases until supersaturation with respect to calcite is obtained in the vesicle and calcium carbonate starts to precipitate at [CO 2− 3 ] V = [CO 2− 3 ] sat .The contribution of bicarbonate conversion to the CO 2− 3 pool in the vesicle is given by: [CO 2− 3 ] sat − f × [CO 2− 3 ] ext (Fig. 2).The H + generated during CO 2− 3 formation from HCO − 3 has to be removed from the vesicle to keep the pH value of the calcifying vesicle constant.Chemical equilibrium in the intracellular carbon dioxide system is obtained after several seconds (approximately 15 s at seawater pH = 8.2 and 25 • C, Zeebe et al., 1999), which is negligible compared with the 1.4 h (calculated using data in Langer et al., 2006) required for the formation of one coccolith (of C. leptoporus).The time to establish δ 18 O equilibrium, on the other hand, is one order of magnitude higher than the time required for coccolith formation (see Table 3.3.9 of Zeebe and Wolf-Gladrow, 2001).It is therefore reasonable to assume that the carbon dioxide system in the calcifying vesicle approaches its chemical equilibrium, whereas the CO 2− It is still a matter of debate which compartment of the cell takes up Ca 2+ from the cytoplasm.It has been suggested that the peripheral endoplasmic reticulum (ER) could play this role (Berry et al., 2002).Using the seawater saturation product (K sp = 10 −6.415 mol −2 kg −2 at S = 32 and T = 20 • C) and for [Ca 2+ ] sat a typical value for [Ca 2+ ] in the ER (500 µmol kg −1 ; Meldolesi and Pozzan, 1998) yields a [CO 2− 3 ] sat in the vesicle of 769 µmol kg −1 .The oxygen isotope fractionation factors (α) between DIC species and water can be calculated using the equations derived by Beck et al. (2005).In freshwater at 20