Phytoplankton calcification as an effective mechanism to alleviate cellular calcium poisoning

Abstract

Marine phytoplankton have developed the remarkable ability to tightly regulate the concentration of free calcium ions in the intracellular cytosol at a level of $\sim$ 0.1 $\mathrm{\mu }$mol L${}^{-\mathrm{1}}$ in the presence of seawater Ca${}^{\mathrm{2}+}$ concentrations of 10 mmol L${}^{-\mathrm{1}}$. The low cytosolic calcium ion concentration is of utmost importance for proper cell signalling function. While the regulatory mechanisms responsible for the tight control of intracellular Ca${}^{\mathrm{2}+}$ concentration are not completely understood, phytoplankton taxonomic groups appear to have evolved different strategies, which may affect their ability to cope with changes in seawater Ca${}^{\mathrm{2}+}$ concentrations in their environment on geological timescales. For example, the Cretaceous (145 to 66 Ma), an era known for the high abundance of coccolithophores and the production of enormous calcium carbonate deposits, exhibited seawater calcium concentrations up to 4 times present-day levels. We show that calcifying coccolithophore species (Emiliania huxleyi, Gephyrocapsa oceanica and Coccolithus braarudii) are able to maintain their relative fitness (in terms of growth rate and photosynthesis) at simulated Cretaceous seawater calcium concentrations, whereas these rates are severely reduced under these conditions in some non-calcareous phytoplankton species (Chaetoceros sp., Ceratoneis closterium andHeterosigma akashiwo). Most notably, this also applies to a non-calcifying strain of E. huxleyi which displays a calcium sensitivity similar to the non-calcareous species. We hypothesize that the process of calcification in coccolithophores provides an efficient mechanism to alleviate cellular calcium poisoning and thereby offered a potential key evolutionary advantage, responsible for the proliferation of coccolithophores during times of high seawater calcium concentrations. The exact function of calcification and the reason behind the highly ornate physical structures of coccoliths remain elusive.

Introduction

Calcium is a versatile and crucial ion in biological systems (Case et al., 2007), which is, among other functions, essential for cellular signalling, membrane structure and cell division (Sanders et al., 1999). The concentrations of cytosolic free Ca${}^{\mathrm{2}+}$ in eukaryotes are well regulated and the maintenance of relatively low levels is essential for fast signal transduction. An excessive influx of Ca${}^{\mathrm{2}+}$ to the cytosol can be lethal as it disturbs intracellular signalling and irreversibly damages the cell (Orrenius et al., 1989; Kader and Lindberg, 2010). Homeostasis of Ca${}^{\mathrm{2}+}$ in plant cells is predominantly achieved by Ca${}^{\mathrm{2}+}$-binding proteins, reducing the effective diffusion coefficient of Ca${}^{\mathrm{2}+}$ in the cytosol, and ultimately via sequestration by the endoplasmic reticulum, mitochondria and cellular vacuoles (Case et al., 2007). Cytosolic free Ca${}^{\mathrm{2}+}$ concentrations in marine phytoplankton are about 10${}^{\mathrm{5}}$ times lower than modern seawater concentrations and marine eukaryotes have developed a remarkable capacity to maintain these low cytosolic Ca${}^{\mathrm{2}+}$ levels (Brownlee et al., 1987, 1995). It is, however, unknown whether the regulating mechanisms of marine phytoplankton to keep this delicate Ca${}^{\mathrm{2}+}$ homeostasis differ between species and between functional groups. In freshwater environments, for example, calcium ions play an important role shaping microalgal species composition. Desmid green algae have a narrow tolerance to calcium (Moss, 1972; Tassigny, 1971) and thrive in soft-water lakes, while submersed macrophytes (Elodea, Stratiotes, Potamogeton) and benthic cyanobacteria dominate in hard-water lakes, where they can be heavily encrusted with CaCO${}_{\mathrm{3}}$ precipitates.

An early hypothesis describes the invention and the process of biomineralization in the form of calcium carbonate by marine organisms as a potential Ca${}^{\mathrm{2}+}$ detoxification mechanism (Simkiss, 1977; Kaźmierczak et al., 1985; Kempe and Degens, 1985). Ocean calcium concentrations have changed remarkably throughout the Phanerozoic eon (past 541 Myr) as documented by fluid inclusions of marine halite (Horita et al., 2002). Over the past 300 Myr, highest seawater Ca${}^{\mathrm{2}+}$ concentrations are documented for the Cretaceous (145 to 66 Ma; Hönisch et al., 2012), known for massive deposition of biogenic calcareous material produced in the pelagic ocean. Calcifying phytoplankton (coccolithophores) are the dominant planktonic calcifiers in the modern ocean and are responsible for up to half the pelagic production of calcium carbonate (Broecker and Clark, 2009). Coccolithophores form minute calcite plates (coccoliths) inside a specialized cell compartment (coccolith vesicle) from where the coccoliths are subsequently transported to the cell's surface and released via exocytosis. The record of nannofossils and coccoliths has its origin in the Late Triassic (about 225 Ma), coinciding with relatively low seawater Ca${}^{\mathrm{2}+}$ concentrations (Bown et al., 2004). Subsequently, seawater Ca${}^{\mathrm{2}+}$ concentrations increased, potentially linked to changes in the seafloor spreading rates (Skelton, 2003), and peaked in the Cretaceous at the highest levels since the past 300 Myr ($\sim$ 3 to 4 times the present seawater concentrations of 10 mmol Ca${}^{\mathrm{2}+}$ L${}^{-\mathrm{1}})$. Species diversity and abundance of total nannofossils, including coccolithophores, have increased in concert with high seawater Ca${}^{\mathrm{2}+}$ concentrations (Fig. 1).

We tested two calcifying coccolithophores (Emiliania huxleyi and Gephyrocapsa oceanica), two diatoms (Chaetoceros sp. and Ceratoneis closterium) and one raphidophyte (Heterosigma akashiwo) to elevated seawater calcium concentrations simulating changes in oceanic Ca${}^{\mathrm{2}+}$ levels over the past 300 Myr. Representative for a non-calcifying coccolithophore, one non-coccolith-carrying (naked) E. huxleyi strain was tested. Furthermore, a possible stimulation of coccolith production by increased seawater Ca${}^{\mathrm{2}+}$ concentration was investigated in two under-calcifying E. huxleyi strains. If biogenic calcification represents a viable mechanism to cope with high external Ca${}^{\mathrm{2}+}$ concentrations, a diverging response in physiological parameters would be expected between calcifiers and non-calcifiers.

Seawater Ca${}^{\mathrm{2}+}$ concentration and fossil phytoplankton diversity over the past 300 Myr. Model-reconstructed seawater Ca${}^{\mathrm{2}+}$ concentration (blue line; data retrieved from Hönisch et al., 2012), fossil species diversity of diatoms (red line; data retrieved from Kooistra et al., 2007), total nannofossils and coccolithophores (black and grey line, respectively; data retrieved from Bown et al., 2004).

[Figure omitted. See PDF]

Materials and methods

Culture conditions

Monospecific cultures of the diploid coccolithophores Gephyrocapsa oceanica (CS-335/03) and Emiliania huxleyi (calcifying CS-370, non-calcifying SO-6.13 and under-calcifying SO-5.25 and SO-8.04), the diatoms Chaetoceros sp. (CHsp-TB02) and Ceratoneis closterium (CCMMG-3), and the raphidophyte Heterosigma akashiwo (CS-169) were grown in sterile artificial seawater (Kester et al., 1967) with macro- and micronutrient additions according to f/2 and f/20 (Guillard, 1975), respectively, or in the case of G. oceanica according to GSe/20 (Loeblich and Smith, 1968). The under-calcified populations (strains SO-5.25 and SO-8.04) consist of cells with no or single attached coccoliths. Cells with no coccoliths attached in these populations either lost their coccoliths, lacked the ability to produce coccoliths or did not yet produce coccoliths. Emiliania huxleyi strain SO-6.13 was isolated by Suellen Cook in February 2007 from the Southern Ocean (54${}^{\circ }$ S, 146${}^{\circ }$ E; 65 m depth). Multiple single-cell isolates from this water sample resulted in a number of calcified ecotype B/C E. huxleyi strains. Strain SO-6.13, however, was naked upon isolation and throughout the conduct of the current study. Much later, in early 2015, strain SO-6.13 switched from a non-calcifying to a calcifying stage and started to produce typical B/C coccoliths.

Calcium concentrations were adjusted by varying additions of CaCl${}_{\mathrm{2}}$ with concomitant additions of NaCl, keeping the ionic strength of the artificial seawater constant. Gephyrocapsa oceanica, H. akashiwo and E. huxleyi (CS-370) were obtained from the Australian National Algae Culture Collection. Ceratoneis closterium was obtained from the Centre of Climate, Meteorology and Global Change at the University of Azores (CMMG). All other species and strains were obtained from the Algae Culture Collection at the Institute of Marine and Antarctic Studies at the University of Tasmania, Australia.

Experimental setup

In the first experiment, cells were acclimated to the experimental conditions (Ca${}^{\mathrm{2}+}$ range from 1 to 52 mmol L${}^{-\mathrm{1}})$ for more than 50 generations and allowed to consume a maximum of 10 % (non-calcifiers) or 5 % (calcifiers) of dissolved inorganic carbon to avoid major changes in the carbonate chemistry. Cultures were incubated in triplicates at 12 ${}^{\circ }$C (16 ${}^{\circ }$C for G. oceanica), a photon flux density of 100 $\mathrm{\mu }$mol quanta m${}^{-\mathrm{2}}$ s${}^{-\mathrm{1}}$ and a $16:\mathrm{8}$ h light : dark cycle at the University of Tasmania. Ceratoneis closterium was incubated at 20 ${}^{\circ }$C, 250 $\mathrm{\mu }$mol quanta m${}^{-\mathrm{2}}$ s${}^{-\mathrm{1}}$ and a $14:10$ h light : dark cycle at the University of Azores. The physiological response of all species (except C. closterium) was examined in terms of growth rate, particulate organic and inorganic carbon cell quota and production rate, and maximum quantum yield of the photosystem II (Fv/Fm). Physiology of C. closterium was only examined in terms of growth rate. Seawater carbonate chemistry was determined from total alkalinity ($A_{\text{T}})$ and dissolved inorganic carbon ($C_{\text{T}})$ samples taken at the start and the end of the experiment.

In the second experiment, two under-calcified E. huxleyi strains (SO-5.25 and SO-8.04) were cultured at the University of Tasmania in triplicates for 2 months under dilute semi-continuous batch conditions at the identical conditions as described above with Ca${}^{\mathrm{2}+}$ concentrations adjusted to 10 or 36 mmol Ca${}^{\mathrm{2}+}$ L${}^{-\mathrm{1}}$. Strain-specific growth rate and the number of coccoliths per cell were monitored over time via cell counts and scanning electron microscopy, respectively. Cultures were allowed to grow from $\sim$ 50 to a maximal cell density of $\sim$ 80 000 cells mL${}^{-\mathrm{1}}$, which prevented major changes in the seawater carbonate chemistry.

Seawater chemistry analysis

Seawater Ca${}^{\mathrm{2}+}$ concentrations at the start of the experiment were determined via chelation ion chromatography (Meléndez et al., 2013), using an adjusted method to match the different Ca${}^{\mathrm{2}+}$ concentrations (precision of $\pm$1.4 %). Dissolved inorganic carbon and $A_{\text{T}}$ were analysed as the mean of triplicate measurements with the infrared detection method using an Apollo SciTech DIC analyser (model AS-C3) and the potentiometric titration method (Dickson et al., 2003), respectively. Data were corrected to certified reference materials (Scripps Institution of Oceanography, USA). Consecutive measurements of the Dickson standard resulted in an average precision of > 99.8 % for both $C_{\text{T}}$ and $A_{\text{T}}$. The carbonate system was calculated using equations from Zeebe and Wolf-Gladrow (2001) with dissociation constants for carbonic acid after Roy et al. (1993), modified with sensitivity parameters for [Na${}^{+}$], [Mg${}^{\mathrm{2}+}$] and [Ca${}^{\mathrm{2}+}$] (Ben-Yaakov and Goldhaber, 1973). The calcite saturation state ($\mathrm{\Omega })$ was calculated with regard to the Mg $/$ Ca ratio as described in Tyrrell and Zeebe (2004). Detailed information on the carbonate system parameters can be found in the Supplement.

Physiological parameters

Maximum quantum yield of the photosystem II (Fv/Fm) was measured on dark-adapted samples (45 min) using a WATER-PAM fluorometer (Walz GmbH, Germany). Subsamples for total particulate carbon (TPC) and particulate organic carbon (POC) were filtered onto pre-combusted (7 h, 450 ${}^{\circ }$C) quartz-microfibre filters (pore-size of 0.3 $\mathrm{\mu }$m) and stored at $-$24 ${}^{\circ }$C. Filters for POC analysis were fumed with saturated HCl for 10 h to remove all inorganic carbon. TPC and POC were measured on an elemental analyser (Thermo Finnigan EA 1112, Central Science Laboratory of the University of Tasmania). Particulate inorganic carbon (PIC) was calculated as the difference between TPC and POC. Cell numbers were obtained by means of triplicate measurements with a Multisizer 4 Coulter Counter (Beckman Coulter, USA) or by light microscopy counts. The average cell number was used to calculate the growth rate $\mathit{\mu }$ (d${}^{-\mathrm{1}})$ as $\mathit{\mu }=$ (ln($c_{\mathrm{1}})$ $-$ ln($c_{\mathrm{0}}\left)\right)$/($t_{\mathrm{1}}$ $-$ $t_{\mathrm{0}})$, where $c_{\mathrm{0}}$ and $c_{\mathrm{1}}$ are the cell concentrations at the beginning ($t_{\mathrm{0}})$ and the end of the incubation period ($t_{\mathrm{1}})$. POC and PIC production rates were calculated from cell quota and species-specific growth rates.

Phytoplankton physiological responses to seawater Ca${}^{\mathrm{2}+}$ concentration. Displayed are laboratory-cultured strains of diatoms (red markers), raphidophytes (blue markers), coccolithophores (black markers) and a non-calcifying coccolithophore (black open marker): (a) species-specific growth rate, (b) maximum quantum yield of photosynthesis (Fv/Fm), (c) cellular POC and (d) PIC quotas, (e) cellular POC and (f) PIC production rates as a function of seawater Ca${}^{\mathrm{2}+}$ concentration. Error bars denote $\pm$1 SD ($n=$ 3). Note that the physiological response of Ceratoneis closterium was only determined via growth rate measurements. POC quota of H. akashiwo could not be determined at a Ca${}^{\mathrm{2}+}$ concentration of 42 mmol L${}^{-\mathrm{1}}$ due to lack of growth.

[Figure omitted. See PDF]

Scanning electron microscopy

Samples for electron microscopy were filtered gently onto polycarbonate filters, air-dried at 60 ${}^{\circ }$C and afterwards sputter-coated with gold–palladium. Photographs were taken with a Hitachi SU-70 field emission scanning electron microscope (SEM) at the Central Science Laboratory of the University of Tasmania. During SEM sessions, > 50 cells were visually evaluated and representative pictures were taken.

Relative physiological response of phytoplankton species to seawater Ca${}^{\mathrm{2}+}$ concentration. Relative fitness expressed in terms of (a) growth rate and (b) POC production of all tested species normalized to ambient seawater Ca${}^{\mathrm{2}+}$ concentration of $\sim$ 10 mmol L${}^{-\mathrm{1}}$, and supplemented with coccolithophore literature data from Müller et al. (2011) and Herfort et al. (2004) to illustrate the effect of calcium poisoning on calcifiers and non-calcifiers. Solid lines indicate regressions through calcifiers: (a) $y=-0.0036\times +1.0483$ ($r^{\mathrm{2}}=$ 0.278, $p=$ 0.035, $n=$ 16) and (b) $y=-0.0052\times +1.0704$ ($r^{\mathrm{2}}=$ 0.184, $p=$ 0.067, $n=$ 19). Dotted lines indicate regressions through non-calcifiers: (a) $y=-0.025\times +1.307$ ($r^{\mathrm{2}}=$ 0.858, $p$ < 0.0001, $n=$ 20) and (b) $y=-0.024\times +1.303$ ($r^{\mathrm{2}}=$ 0.826, $p$ < 0.0001, $n=$ 15).

[Figure omitted. See PDF]

Representative SEM photographs of the under-calcified E. huxleyi strain SO-8.04 cultured at modern seawater Ca${}^{\mathrm{2}+}$ concentration of 10 mmol L${}^{-\mathrm{1}}$, showing no or only single attached coccoliths (a). When cultured for 2 months at elevated Ca${}^{\mathrm{2}+}$ concentration of 36 mmol Ca${}^{\mathrm{2}+}$ kg${}^{-\mathrm{1}}$, E. huxleyi strain SO-8.04 produced a sufficient number of coccoliths to cover the whole cell (b).

[Figure omitted. See PDF]

Results

In the first experiment, at Ca${}^{\mathrm{2}+}$ concentrations below 2 mmol L${}^{-\mathrm{1}}$, all species exhibited significantly ($t$ test, $p$ < 0.05) lower growth, particulate organic carbon (POC) production rates and maximum quantum yield of photosystem II (Fv/Fm) compared to modern seawater concentrations of $\sim$ 10 mmol Ca${}^{\mathrm{2}+}$ L${}^{-\mathrm{1}}$ (Fig. 2). Furthermore, the two calcifying species displayed decreased particulate inorganic carbon (PIC) production rates at Ca${}^{\mathrm{2}+}$ concentrations below 2 mmol L${}^{-\mathrm{1}}$ compared to $\sim$ 10 mmol Ca${}^{\mathrm{2}+}$ L${}^{-\mathrm{1}}$ ($t$ test, $p$ < 0.05). At elevated Ca${}^{\mathrm{2}+}$ concentrations all non-calcifying species exhibited a severe reduction in growth, POC production and maximum quantum yield (Fig. 2). In the most extreme cases no growth was detected at 42 and 52 mmol Ca${}^{\mathrm{2}+}$ L${}^{-\mathrm{1}}$ in H. akashiwo and C. closterium, respectively. Both tested coccolithophore species, on the other hand, were able to maintain their growth, Fv/Fm, POC and PIC production rates with no substantial change at calcium concentration expected for Cretaceous seawater (25 to 40 mmol Ca${}^{\mathrm{2}+}$ L${}^{-\mathrm{1}})$. A further increase in external Ca${}^{\mathrm{2}+}$ concentrations up to 52 mmol L${}^{-\mathrm{1}}$ adversely affected POC and PIC production only in E. huxleyi, whereas G. oceanica was not impaired. The non-calcifying strain of E. huxleyi exhibited a similar response to that of the diatom and raphidophyte species with reduced physiological rates of up to 84 % at Ca${}^{\mathrm{2}+}$ concentrations of 19 mmol L${}^{-\mathrm{1}}$ and higher (Fig. 2). To illustrate the diverging physiological response of calcifying coccolithophores and non-calcifying phytoplankton, we normalized growth and POC production rates from the current study and literature data to the species-specific rates exhibited at modern ocean calcium levels (Fig. 3). A linear regression fit (from 9 to 52 mmol Ca${}^{\mathrm{2}+}$ L${}^{-\mathrm{1}})$ through calcifiers and non-calcifiers resulted in a 6.9 times steeper reduction for the latter group in terms of growth rate (Fig. 3a) and a 4.6 times steeper reduction in terms of POC production rates (Fig. 3b).

In the second experiment, the two under-calcified E. huxleyi strains (SO-5.25 and SO-8.04) cultured at elevated seawater Ca${}^{\mathrm{2}+}$ concentrations (36 mmol L${}^{-\mathrm{1}})$ displayed no significant change in growth rate ($t$ test, $p$ > 0.05) compared to strains cultured at modern Ca${}^{\mathrm{2}+}$ concentrations of 10 mmol L${}^{-\mathrm{1}}$ (0.67 $\pm$ 0.01 and 0.72 $\pm$ 0.01 d${}^{-\mathrm{1}}$ compared to 0.68 $\pm$ 0.01 and 0.71 $\pm$ 0.01 d${}^{-\mathrm{1}}$ for the strains SO-5.25 and SO-8.04, respectively). The number of coccoliths per cell, however, increased remarkably from fewer than 2 coccoliths per cell at 10 mmol Ca${}^{\mathrm{2}+}$ L${}^{-\mathrm{1}}$ to more than 12 coccoliths per cell, forming a complete coccosphere, at 36 mmol Ca${}^{\mathrm{2}+}$ L${}^{-\mathrm{1}}$ (Fig. 4).

Discussion

The results presented here demonstrate the influence of seawater Ca${}^{\mathrm{2}+}$ concentrations on marine phytoplankton physiology (in terms of growth and particulate organic carbon production). Whereas previous studies have already investigated the effects of elevated seawater Ca${}^{\mathrm{2}+}$ concentrations on calcifying coccolithophore physiology and coccolith formation (Herfort et al., 2004; Langer et al., 2007; Müller et al., 2011), this study is to our knowledge the first to investigate the Ca${}^{\mathrm{2}+}$ sensitivity of non-calcifying phytoplankton in the laboratory. Marine phytoplankton presumably operate several mechanisms which contribute to cellular Ca${}^{\mathrm{2}+}$ regulation, such as intra- and extracellular enzymatic binding capacities and/or the influx regulation via selective channels (Gadd, 2010). Over the past decade progress has been made in the discovery of cellular compartments (e.g. endoplasmic reticulum, chloroplast, mitochondria) regulating plant Ca homeostasis and signalling (McAinsh and Pittmann, 2009; Webb, 2008; Brownlee and Hetherington, 2011), as well as in the differences between the Ca channels of eukaryotes, higher plants and mammalian cells (Wheeler and Brownlee, 2008). However, many unknowns remain about phytoplankton intracellular ion regulation and the homeostasis of the major biological active cations like Ca${}^{\mathrm{2}+}$ and Mg${}^{\mathrm{2}+}$ and their interaction and possible influence on each other. For example, Ca${}^{\mathrm{2}+}$ has a higher ion-exchange capacity than Mg${}^{\mathrm{2}+}$ (Harris, 2010) and when present in high concentrations might interfere with enzymatic reactions where Mg${}^{\mathrm{2}+}$ acts as a cofactor (Moore et al., 1960; Legong et al., 2001). However, it remains speculative whether this is a possible explanation for the observed reduction in growth rate and Fv/Fm of non-calcifying phytoplankton species (Fig. 2).

The non-calcifying strain of E. huxleyi showed a comparable response to elevated seawater Ca${}^{\mathrm{2}+}$ concentrations as the diatom and raphidophyte species (Fig. 3). This indicates that the Ca${}^{\mathrm{2}+}$ tolerance of calcifying coccolithophores compared to non-calcifying phytoplankton is not a taxon-specific trait but connected to the process of calcification itself and, furthermore, suggests that coccolithophore biomineralization acts as an efficient mechanism to cope with high external Ca${}^{\mathrm{2}+}$ concentrations. Reduced overall fitness triggered by high external Ca${}^{\mathrm{2}+}$ concentrations is presumably associated with enhanced transmembrane Ca${}^{\mathrm{2}+}$ influx, leading to higher energetic costs for cytosolic Ca${}^{\mathrm{2}+}$ removal and might ultimately result in a disadvantage in resource competition between phytoplankton species. Dunaliella, a member of the class Chlorophyceae, is one of the most tolerant phytoplankton species regarding high external ion concentrations and regularly blooms in highly saline lakes (Oren, 2002, 2005). However, this extremophile species is inhibited in growth by high external Ca${}^{\mathrm{2}+}$ concentrations and only forms blooms in hypersaline lakes when the upper water layer becomes sufficiently diluted with regard to its Ca${}^{\mathrm{2}+}$ concentrations (Baas-Becking, 1931). This emphasizes the ecological importance of external Ca${}^{\mathrm{2}+}$ concentrations for phytoplankton growth dynamics.

The remarkable tolerance of calcifying coccolithophores to elevated Ca${}^{\mathrm{2}+}$ concentrations likely results from a tight control on transmembrane Ca${}^{\mathrm{2}+}$ entry, intracellular transport, and deposition. Seawater Ca${}^{\mathrm{2}+}$ presumably enters the coccolithophore cell through permeable channels into the peripheral endoplasmatic reticulum. Via the endomembrane transport network it reaches a Golgi-derived organelle, the coccolith vesicle, where it is precipitated as CaCO${}_{\mathrm{3}}$ (Mackinder et al., 2010). Precipitation of Ca${}^{\mathrm{2}+}$ in the form of calcite changes the ion to a biochemically inert state. Large amounts of Ca${}^{\mathrm{2}+}$ can thereby be sequestered in a finite space and time. For Emiliania huxleyi to sustain a typical rate of calcification, an uptake of 5 $\times$ 10${}^{\mathrm{6}}$ Ca${}^{\mathrm{2}+}$ ions s${}^{-\mathrm{1}}$ is required (Mackinder et al., 2010). The fact that this massive intracellular Ca${}^{\mathrm{2}+}$ flux needs to be achieved at a cytosolic concentration of only 100 nmol Ca${}^{\mathrm{2}+}$ L${}^{-\mathrm{1}}$ without disturbing the cell's delicate Ca${}^{\mathrm{2}+}$ homeostasis exemplifies the level of cellular control involved in coccolithophore calcification. It appears reasonable to assume that this tight cellular control of biogenic calcification (which includes CaCO${}_{\mathrm{3}}$ precipitation inside the coccolith vesicle and the regulation of cellular Ca${}^{\mathrm{2}+}$ entrance and distribution) also allows for the observed tolerance to external Ca${}^{\mathrm{2}+}$ concentrations. The absence of Ca${}^{\mathrm{2}+}$-stimulated calcification at levels above modern ocean Ca${}^{\mathrm{2}+}$ concentrations (Fig. 2f) is in line with previous findings, which indicate saturation of calcification in E. huxleyi and C. braarudii at $\sim$ 10 mmol Ca${}^{\mathrm{2}+}$ L${}^{-\mathrm{1}}$ (Herfort et al., 2004; Trimborn et al., 2007; Leonardos et al., 2009; Müller et al., 2011). This suggests that in coccolithophores adapted to modern ocean conditions, factors other than the Ca${}^{\mathrm{2}+}$ concentration may limit CaCO${}_{\mathrm{3}}$ precipitation at higher than ambient Ca${}^{\mathrm{2}+}$ levels. Potentially limiting factors include dissolved inorganic carbon acquisition and energy supply for the process of calcification (Bolton and Stoll, 2013; Bach et al., 2015).

Emiliania huxleyi is characterized by three distinct different cell forms: (a) the coccolith-carrying non-motile diploid form (C cell), (b) the naked non-motile diploid form (N cell) and (c) the scaly motile haploid form (S cell). The latter haploid form possesses organic body scales covering the cell and two flagellates that enable motion (Paasche, 2002). The life cycle of E. huxleyi consists of C and S cells, whereas N cells are mostly observed in the laboratory after extended culture periods (Paasche, 2002) or under unfavourable culture conditions (Müller et al., 2015). This study investigated only the diploid coccolith-carrying (C cell) and the naked (N cell) cell forms of E. huxleyi. Our observations and the presence of N and S cells in laboratory cultures and natural populations (Paasche, 2002; Frada et al., 2012; Müller et al., 2015) indicate that E. huxleyi cells have the ability to control intracellular Ca${}^{\mathrm{2}+}$ homeostasis at modern Ca${}^{\mathrm{2}+}$ concentrations without the need of biomineralization.

At modern seawater conditions some E. huxleyi strains display an incomplete coccolith cover (coccosphere) with less than 2 coccoliths per cell (Fig. 4a) instead of the 10 to 15 that are necessary to form a complete coccosphere (Paasche, 2002). The results of the second experiment indicate that an existent but under-saturated calcification mechanism can be stimulated by increased seawater Ca${}^{\mathrm{2}+}$ concentrations (Fig. 4b) and, furthermore, might prevent cellular Ca${}^{\mathrm{2}+}$ poisoning as seen in the non-calcifying E. huxleyi strain (Figs. 2 and 3). However, benefits of coccolith formation are expected which evidently outweigh the substantial costs of this energy-consuming process even under modern ocean Ca${}^{\mathrm{2}+}$ concentrations. Although numerous hypotheses have been proposed concerning the precise function of coccolithophore calcification, including ballasting and protection from viruses, grazers and damaging irradiance, so far none of these is conclusively supported by experimental evidence (Raven and Crawfurd, 2012; Barcelos e Ramos et al., 2012).

Palaeoecological implications

Palaeoceanographic studies have indicated that the oceanic conditions of the Cretaceous were quite different from those in the modern ocean (e.g. see Zeebe, 2001; Hay, 2008). Besides elevated seawater Ca${}^{\mathrm{2}+}$ concentrations (Fig. 1), the Cretaceous was marked by a warm greenhouse environment, elevated sea levels, warm shallow shelf seas and altered oceanic circulation. Here we tested whether the biomineralization mechanism in coccolithophores increases their resilience to cellular calcium stress, which indeed is indicated by the physiologically different responses of the three calcifying coccolithophore species (E. huxleyi, G. oceanica and C. braarudii) compared to the non-calcifying species (Fig. 3). Cretaceous seawater Ca${}^{\mathrm{2}+}$ concentrations may thus have represented a selective advantage for coccolithophores during this period of the geological past. This could explain the proliferation and high productivity of coccolithophores during the Cretaceous compared to non-calcifying phytoplankton. We cannot exclude the possibility of other environmental factors that might have supported the proliferation of coccolithophores or suppressed non-calcifiers in the Cretaceous (e.g. Stanley et al., 2005), but the seawater Ca${}^{\mathrm{2}+}$ concentrations seem to be a major environmental aspect promoting coccolithophore over non-calcifying phytoplankton growth.

It remains an open question whether the onset of calcification in coccolithophores (approx. 225 Ma) at relatively low seawater Ca${}^{\mathrm{2}+}$ concentrations evolved primarily to efficiently regulate cellular Ca${}^{\mathrm{2}+}$ homeostasis or whether calcification had other functions at that time. If calcification in coccolithophores evolved as a Ca${}^{\mathrm{2}+}$ detoxification mechanism, it was presumably an additional instrument to regulate intracellular Ca${}^{\mathrm{2}+}$ levels because other strategies must have existed in the ancestors of coccolithophores that did not precipitate calcium carbonate. It is reasonable to assume that the rising oceanic Ca${}^{\mathrm{2}+}$ concentrations represented a selective pressure on phytoplankton populations and may have provided an evolutionary advantage to coccolithophores over non-calcareous phytoplankton during the Jurassic and Cretaceous period (Fig. 1). However, secondary benefits of calcification are likely responsible for its continued operation under modern ocean Ca${}^{\mathrm{2}+}$ concentrations. Interestingly, E. huxleyi and G. oceanica, the dominant coccolithophores in the modern ocean, are two of the few coccolithophore species that have a non-calcifying haploid life stage, whereas the haploid life stage of the majority of coccolithophores is calcified (Billard and Inouye, 2004). This led us to suggest that these two species in the modern ocean do not rely on cellular Ca${}^{\mathrm{2}+}$ detoxification by biomineralization.

Concluding remarks

The concept of biocalcification as a Ca${}^{\mathrm{2}+}$ detoxification mechanism in marine organisms has been proposed earlier (Simkiss, 1977; Kaźmierczak et al., 1985) and, based on the results of this study, is supported for coccolithophores. The occurrence of calcified cyanobacteria in the geological record during the Phanerozoic also appears to be connected to elevated seawater Ca${}^{\mathrm{2}+}$ concentrations (Arp et al., 2001), suggesting similarities in the benefits of calcification in fossil cyanobacteria and coccolithophores. It remains speculative to extend the “Ca${}^{\mathrm{2}+}$-detoxification concept” to other marine calcifying groups or to the onset of biocalcification in the Precambrian–Cambrian transition (Kempe and Kaźmierczak, 1994; Brennan et al., 2004). However, in view of the substantial variability in seawater Ca${}^{\mathrm{2}+}$ concentration during Earth's history and the observed Ca${}^{\mathrm{2}+}$ sensitivity of dominant marine phytoplankton species, the ocean's Ca${}^{\mathrm{2}+}$ ion concentration should be considered a potential factor influencing the evolution of marine life on Earth.

The Supplement related to this article is available online at doi:10.5194/bg-12-6493-2015-supplement.

Acknowledgements

We thank D. Davis for laboratory assistance and A. McMinn for providing a PAM fluorometer. We are grateful for the constructive comments of T. Tyrrell, J. Young and one anonymous reviewer. Additional comments from a research group meeting (composed of L. Munns, M. Duret, C. Daniels, K. Mayers, A. Poulton and R. Sheward) further increased the quality of the manuscript. The work was funded by the Australian Research Council (DP 1093801 to G. M. Hallegraeff and T. W. Trull) and the “Conselho Nacional de Desenvolvimento Científico e Tecnológico Brasil (CNPq, Processo: 405585/2013-6)”. K. G. Schulz is the recipient of an Australian Research Council Future Fellowship (FT120100384). We thank the data publisher PANGAEA for hosting and making the data fully available under http://doi.pangaea.de/10.1594/PANGAEA.854719. Edited by: J.-P. Gattuso