The Mg/Ca ratio in foraminiferal shells is widely used as a proxy for paleotemperatures. Nevertheless, it seems that the basic Mg content of foraminifera is determined by biological factors, as can be concluded from the large inter species and intrashell variability and the frequent deviations from inorganic behavior. This paper discusses three possible ways by which foraminifera can control or modify the Mg content in their shell: (1) involvement of organic matrix in the precipitation process that may alter the partition coefficient of Mg in biogenic calcite, (2) controlled conversion of transient amorphous phases to calcite, and (3) modification of the Mg concentration in the parent solution from which the crystals precipitate. The first two mechanisms are probably responsible for the precipitation of high-Mg calcite phases (whole shell or sublayers), while the third mechanism leads to the formation of low-Mg calcite phases. We propose a model adapted from epithelial cells that allows massive Mg2+ removal from the biomineralization site. This model is especially relevant to the planktonic and deep benthic low-Mg foraminifera that are frequently used for paleotemperature reconstructions. We discuss the possible biological roles of Mg in the shell in terms of the calcite polymorph conservation, the in vivo chemical stability of the shell, the functions of Mg as a stabilizer of transient phases and as a controlling agent of the precipitation process. Several temperature sensitive biological processes that may influence the Mg/Ca ratio of the shell are suggested and a model that combines biogenic and inorganic considerations is presented. The model uses Mg heterogeneity in the shell together with temperature response (biologic and inorganic) of biomineralization processes, to account for the deviation of planktonic foraminifera from inorganic calcite at equilibrium with seawater.
Field and laboratory calibrations demonstrated that Mg/Ca variations within a foraminiferal species are well correlated with temperature changes [e.g., Nurnberg et al., 1996; Rosenthal et al., 1997; Lear et al., 2000; Anand et al., 2003; Russell et al., 2004]. However, it was recognized that there are major “vital effects” that influence the initial Mg/Ca ratio of a given species and its response to temperature. The physiological processes that govern and/or influence the Mg content in foraminifera shells and their response to temperature are poorly known. We will try to elucidate this issue and consider some potential physiological-mediated responses of the Mg/Ca ratio to temperature in foraminifera.
1.1. Taxonomic Variability of Mg/Ca Ratio
The foraminifera as a group display a very large variability in the Mg content of their carbonate shells, ranging between more than 20 mole% to less than 0.1 mole% of MgCO3 [Blackmon and Todd, 1959] (Figure 1). On the basis of Mg content, the foraminifera have been divided to three major groups: high-Mg (>10%), intermediate-Mg (5–10%), and low-Mg (<5%) [Blackmon and Todd, 1959]. Frequently, foraminifera that share the same habitat display significant differences in their Mg concentration, indicating that the Mg content is predominantly determined by biological factors [Blackmon and Todd, 1959]. This is also apparent from the observed intrashell variability [Duckworth, 1977; Brown and Elderfield, 1996; Erez, 2003; Eggins et al., 2004], which cannot be ascribed solely to environmental changes. Blackmon and Todd  concluded that the Mg variability in foraminifera shells is related to their taxonomy and phylogenetic relationships. The planktonic foraminifera are members of the low-Mg group, the Miliolids generally belong to the high-Mg group while the Rotaliids show very large variability among different species within the family (Figure 1). The taxonomically related Mg variability in foraminifera probably represents different genetically controlled biomineralization processes that developed during the course of evolution.
The Mg content in other calcified marine invertebrates (e.g., echinoderms, mollusks, calcareous algae, crustaceans, tube worms, etc.) also shows large taxonomic variability [Chave, 1954; Lowenstam and Weiner, 1989] suggesting a tight biological control on the Mg content. Despite this biological control, these organisms, like the foraminifera, still record in their shells, environmental signals, such as the Mg/Ca ratio of the seawater and temperature [Chave, 1954; Dickson, 2002; Ries, 2004].
1.2. The Parent Solution for Calcification
The cellular control on Mg content in foraminiferal calcite can be exerted in several ways. A major factor influencing the Mg content in foraminifera is the source of the parent solution from which the calcite precipitates. Foraminifera have two basically different optional sources for this solution: ambient seawater or cell-derived fluids; each one bearing its advantages and disadvantages. The advantage of using ambient seawater in today's ocean is that the cell starts with a supersaturated solution (in shallow water) that would require very low additional energy input. However, in order to control the polymorph and the composition of the mineral phase, the cell has to modify the solution by various transport systems. The advantage of using cell-derived fluids is that the composition of the calcifying solution is determined predominantly by the cell. However, in order to achieve a supersaturated solution the cell must expend energy to concentrate the essential ions. This would require a large energy expenditure with regard to Ca2+ since its normal intracellular concentration is <1 μM, and the cell will need to transfer large amounts of Ca2+ to the parent solution while maintaining this extremely low intracellular concentration. The control over the source of the parent solution may differ between foraminifera groups and depends on their biomineralization mechanism.
The biomineralization process in foraminifera is under tight biological control as can be inferred from the complex shape, structure, texture, crystallography, and chemical and isotopic compositions of their shells [e.g., Towe and Cifelli, 1967; Hansen et al., 1969; Lowenstam and Weiner, 1989]. One condition for biologically controlled calcification is the creation of a privileged space in which the ions concentration can be regulated. In foraminifera, two types of privileged spaces that correspond to two calcification modes were identified. The first mode was observed in the Miliolid group (porcelaneous, imperforate) that precipitate high-Mg calcite. In this group, small calcite crystals are precipitated within intracellular vesicles; during new chamber formation these vesicles are exocytosed and the crystals are assembled at the site of chamber formation [Angell, 1980; Hemleben et al., 1986]. The second mode is precipitation in situ, which occurs in an extracellular confined space that is delineated by membranous structures that serve as an isolating barrier [Angell, 1979; Hemleben et al., 1986]. This mode of calcification is used by the calcitic-radial foraminifera that dominate the oceans today. The calcitic-radial foraminifera are hyaline, perforated, lamellar, and display a large range of Mg/Ca ratios. In the first mode (imperforated foraminifera), the chemical composition within the small vesicles located in the highly controlled cellular environment may be tightly regulated and modified by the organism (thus represent cell-derived fluids). The second mode (perforated foraminifera) requires an adequate extracellular parent solution, where the calcification will take place. The control of the parent solution in this case seems to be more complicated and its initial composition is more critical.
On the basis of the observation that perforate foraminifera shells properly record the ambient chemistry and isotopic composition of the seawater in which they live, it was assumed that the parent solution for calcification is close to seawater [Elderfield et al., 1996]. Our new in vivo observations on the calcification process of the perforate species A. lobifera [Bentov et al., 2001] support this assumption. We have shown that seawater is internalized by the cell into large intracellular vacuoles. The vacuolated seawater undergoes chemical modification, which includes alkalization and possibly water removal, as well as other yet undetermined processes. The seawater vacuoles are introduced to the extracellular calcification space where their content serves as the parent solution from which calcification occurs. If this method is universal among the perforated foraminifera, it may imply that in this group the initial Mg/Ca ratio of the calcifying solution is very high (∼5.2 in seawater).
The most important factor that influences the Mg/Ca ratio of inorganic calcite is the Mg/Ca ratio in the parent solution [Mucci and Morse, 1983; Hartley and Mucci, 1996]. Inorganic calcite precipitating from seawater contains 6 and 8% MgCO3 at 5 and 25°C, respectively [Mucci, 1987]. Foraminifera, however, frequently display Mg content that is very different from that of inorganic calcite. Shallow benthic species (both perforate and imperforate) often show high Mg content, exceeding that of inorganic calcite by a factor of up to 2.5 (Figure 1). On the other hand, planktonic and deep benthic species show a large decrease in Mg/Ca relative to inorganic calcite by a factor of up to 70, with values as low as 0.1 mole% [Duckworth, 1977; Nurnberg et al., 1996; Rosenthal et al., 1997]. Assuming that seawater is the parent solution in all perforated foraminifera, it is plausible that the low-Mg foraminifera decrease the Mg/Ca ratio of the parent solution from which the crystals precipitate.
2. Possible Mechanisms to Control Mg/Ca Ratio in the Shell
2.1. Control of the Composition of the Parent Solution
2.1.1. Mg2+ Regulation in Eukaryotic Cells
Mg2+ is unique among biological cations because of its high charge density and its solution chemistry that is related to its large hydration sphere. This is abundantly reflected in its transport systems [Smith and Maguire, 1998]. Magnesium is the most abundant divalent ion within living cells and the second most abundant cellular cation after potassium. It is a cofactor in many enzymes, and its free ionic concentration regulates many metabolic processes and membrane ionic channels. This implies that the cell maintains an internal free Mg2+ concentration, [Mg2+]i, within a narrow defined range as part of its overall homeostasis. The reported [Mg2+]i in eukaryotic cells is between 0.2 to 1.2 mM [Romani and Maguire, 2002], however the total Mg concentration is much higher and can reach 25 mM. This is because Mg may be bound to various organic molecules or sequestered within different compartments of the cell (see below). The eukaryotic cell faces an inward driving force for Mg2+ from the extracellular environment into the cell. This force is composed of two components, the chemical concentration gradient and the electric potential of the membrane. Since eukaryotic cells maintain a membrane potential of −40 to −70 mV (negative inside) by active ions transport (mainly K+ and Na+), there is an inward electric driving force for extracellular cations which is usually much larger than the chemical component of the concentration gradient. The membrane potential and [Mg2+]i have not yet been measured in foraminifera, however it is probable that it is not significantly different from other eukaryotic cells. Assuming a seawater Mg2+ concentration ([Mg2+]o) of 53 mM, a membrane potential −40 mV, and temperature of 25°C, if Mg2+ were allowed to distribute itself at equilibrium, [Mg2+]i would be more than 1 M according to the Nernst equation.
where R is the gas constant, T is the temperature (Kelvin scale), F is the Farady constant, and Z is the ion charge. This is a theoretical calculation only, but it demonstrates the potential of cells to absorb large amounts of Mg2+ from the extracellular fluid. Thus, as other eukaryotes, foraminifera must actively keep a low intracellular Mg2+ activity against a large electrochemical gradient. This may be accomplished by employing three basic methods: Mg transport systems, cellular buffering and sequestration within cellular organelles.
2.1.2. Mg Transport
Several transport systems of Mg2+ were identified in eukaryotic cells: (1) channels that selectively allow-passive Mg2+ diffusion according to the concentration gradient across the cell membrane [Preston, 1998], (2) pumps that transport ions actively against the electrochemical gradient by using the chemical energy of ATP hydrolysis, (3) exchangers that utilize the gradient of one ion to countertransport a different ion with the same charge against its chemical gradient, (4) cotransporters that use the electrochemical gradient of one ion to transport with it a different ion with the opposite charge. All of these transport systems were found for Mg2+. The most abundant Mg2+ transport system is a Na+/Mg2+ exchanger that utilizes the Na+ gradient (low inside) across the membrane to extrude Mg2+ in exchange for two Na+. A Ca2+/Mg2+ exchanger that would be very potent in altering the Mg/Ca ratio in the parent solution has been also found in eukaryotic (mammalian) cells [Cefaratti et al., 1998]. An exchanger of H+/Mg2+ (in plants [Shaul et al., 1999]), a cotransporter of Mg2+/Cl− (in erythrocytes [Gunther et al., 1990]) and a Mg-ATPase pump were also reported [Romani and Maguire, 2002]. It is noteworthy that in addition to the above Mg2+ transport mechanisms the cell can further decrease the Mg/Ca ratio in the parent solution by increasing its Ca2+ concentration using a Ca-ATPase pump as was suggested in corals [Al-Horani et al., 2003].
2.1.3. Mg Buffering
In all cells examined so far [Mg2+]i represents only a small part of the total Mg2+ concentration. In mammalian cells, [Mg2+]i is 0.25–1 mM, whereas the total cellular Mg2+ concentration is between 14 to 20 mM [Romani et al., 1993]. The majority of Mg2+ appears to be bound to phospholipids, proteins, chromatin, nucleic acids and nucleotides, which taken together represent the intracellular Mg2+ buffering system. ATP constitutes one of the major binding molecules for cellular Mg2+ (Mg2+ + ATP4− ↔ MgATP2−) because of its relatively high affinity for Mg2+ and its high intracellular concentration (around 5 mM) [Romani and Maguire, 2002]. ATP, which is derived from mitochondrial oxidative phosphorylation, is the primary utilizable source of high-energy phosphate bonds within the cell; increasing its synthesis will increase the buffering capacity of [Mg2+]i.
Like other cations, [Mg2+]i can be controlled by sequestration into cellular compartments that absorb or release the cation upon request. Mitochondria, together with the endoplasmic reticulum, seem to be the major compartments that serve this purpose [Grubbs et al., 1984]. It is noteworthy that concentration of mitochondria at a specific site will locally increase the buffer capacity for Mg2+ both as a sequestration site and as a major source of ATP. On the basis of TEM observations in foraminifera it was reported that mitochondria concentrate around the calcification site [Hemleben et al., 1986]. Our own TEM and confocal fluorescence imaging of calcifying foraminifera also show high concentration of mitochondria near the calcification site, suggesting a possible role for mitochondria in controlling the Mg/Ca ratio during calcification.
2.2. Possible Influence of the Organic Matrix on Mg Content in the Shell
In chemical precipitation experiments, organic additives such as acidic polysaccharides and carboxylic acids, favor the formation of calcite instead of aragonite [Wada et al., 1999]. They showed that without these additives, aragonite would have been favored because of the presence of Mg2+. The formation of magnesian calcite is explained by the following process: initially aragonite nuclei are formed due to the presence of Mg2+, however the carboxylic acids selectively inhibit additional aragonite growth causing the original aragonite seeds to transform into calcite nuclei through a solid-to-solid transition process. It was further shown that an increase in the amount of the carboxylic acids result in a higher substitution of Mg for Ca in the calcite structure. For example, increasing the concentration of succinic acid from 0.1 mg/10 mL to 10 mg/10 mL caused MgCO3 to increase from 1.5 to 6 mole%, in the calcite [Wada et al., 1999]. It is possible that such organic additives, found in the organic matrix of calcifying invertebrates, increase the Mg content in foraminiferal calcite.
The organic matrix can also influence the Mg/Ca ratio in calcite by favoring growth on preferred faces, since different crystallographic faces show differential partitioning of Mg [Davis et al., 2004]. By controlling the growing crystal face using organic matrix components [Lowenstam and Weiner, 1989], the foraminifera may further control the Mg2+ content in the shell.
The organic matrix organization within the calcareous walls of perforate foraminifera which display laminated patterns [Reiss, 1957] may explain part of the intrashell Mg/Ca heterogeneity. Indeed, in A. lobifera it was observed that the high-Mg calcite phases are associated with layers that are enriched with organic matrix [Erez, 2003]. Furthermore, recently we observed that the first calcitic layer deposited during new chamber formation in A. lobifera is composed of high-Mg calcite [Bentov and Erez, 2005]. This initial layer is strongly associated with the organic template (often termed Primary Organic Membrane, POM [Angell, 1967; Hemleben et al., 1986]. Other reported results on Mg heterogeneity in planktonic foraminifera shells [Duckworth, 1977; Eggins et al., 2003, 2004] may also be associated with the distribution of organic matrix.
2.3. Precipitation From an Amorphous Transient Phase
The formation of calcite and aragonite from transient phases of amorphous CaCO3 (ACC) was demonstrated in echinoderms and mollusks [Beniash et al., 1997; Addadi et al., 2003]. ACC is a metastable phase that normally transforms within minutes into a stable crystalline phase. In biomineralization systems, this transformation is gradual and is probably mediated by the cells [Beniash et al., 1997; Addadi et al., 2003]. If this transformation process is present in foraminifera it may induce various vital effects, including on the Mg/Ca ratio, and may explain the formation of high-Mg calcite in some species. Inorganic precipitation experiments from solutions that are similar to seawater showed that formation of high Mg calcite (up to 21 mol%) is enabled through initial ACC formation that transformed later into calcite [Raz et al., 2000]. Otherwise, without this amorphous intermediate, such high-Mg calcite would have not been precipitated from a similar solution.
The transient amorphous phases may contain variable concentrations of trace and minor elements (including Mg) that are very different from that expected for calcite or aragonite precipitating from seawater. During the crystallization process, differential distribution of Mg may be observed because of segregation processes [Raz et al., 2000]. In calcifying organisms this process may be mediated by the cell.
2.4. The pH or (CO32−) Control
Theoretically higher supersaturation values in the privileged space induced by the organism (e.g., by pH elevation) may allow precipitation of the less stable high Mg calcite phase. Zeebe and Sanyal  have shown in inorganic precipitation experiments, that at high pH, significant rates of CaCO3 precipitation occurred even at high Mg concentrations. In contrast, Russell et al.  showed a slight increase in Mg with a decrease in CO32− concentration in cultured planktonic foraminifera. Carbonate chemistry control on Mg content needs to be further explored both in field and laboratory studies.
3. Temperature Related Biological Processes That May Affect the Mg Content of the Foraminiferal Shell
Although the basic Mg content in foraminifera is determined biologically, it has been widely demonstrated that the Mg/Ca ratio in a given species is responsive to environmental variables, mainly temperature but also salinity and the Mg/Ca ratio in the ambient seawater [Nurnberg et al., 1996; Rosenthal et al., 1997; Lea et al., 1999; E. Segev and J. Erez, Effect of Mg/Ca ratio in seawater on shell composition in shallow benthic foraminifera, submitted to Geochemistry, Geophysics, Geosystems, 2005]. The increase in Mg/Ca ratio with temperature was demonstrated experimentally for inorganic CaCO3 [Katz, 1973; Oomori et al., 1987]. However, laboratory and field calibrations show that the temperature response of the Mg/Ca ratio in low-Mg foraminifera is higher than the inorganic response (reviewed in the work of Lea ). It was speculated that temperature related physiological processes may have a dominant role in regulating the coprecipitation of Mg in these foraminifera [Rosenthal et al., 1997]. It is noteworthy that the increase in the Mg/Ca ratio with temperature in low-Mg foraminifera is significant percentage wise (∼9–10% °C−1). However, the actual Mg increase is negligible compared to the (presumed) substantial Mg2+ removal from the parent solution, assuming it is originally seawater. Since the main cellular effort is probably directed to decrease the Mg/Ca ratio in the shell, it seems that the increase with temperature is a secondary effect. The highly depleted parent solution may be sensitive to even small variations in Mg metabolism, including temperature-dependent variations. These variations may, in turn, induce large changes in the Mg/Ca ratio in the low-Mg shell and explain the augmentation of the Mg temperature response in this group. Indeed, experiments and observations on the Mg/Ca response to temperature in shallow benthic high-Mg foraminifera (presumably with a high-Mg2+ parent solution), showed a moderate response of ∼ 3% °C−1 [Toyofuku et al., 2000] similar to that of inorganic calcite. Because there are many physiological processes that can influence the Mg2+ concentration, and many of them may be temperature sensitive, it is difficult to review all these potential biological effects. Below we present two examples of such possible influences.
3.1. Enhanced ATP Hydrolysis
A common parameter that characterizes enzymatic activity is Q10, which represents the change in activity with a change of 10°C. It is well known that ATP utilization rate increases with temperature. The Q10 for ATP hydrolysis in rabbit muscle, for example, is 2.5 in the temperature range 7°–25°C, but increases to 9.7 in the range 25°–35°C, [Hilber et al., 2001]. As mentioned before, ATP represents one of the major buffering molecules that binds free Mg2+. Therefore any process that increases ATP hydrolysis would result in reduced buffer capacity that may, in turn, cause a higher Mg/Ca ratio in the shell. In this respect, it should be noticed that the Mg2+ buffer capacity of ATP is also pH (intracellular) dependent. In solution, various protonated forms of ATP constitute a complex multiple-equilibrium mixture, with each species having different affinities for Mg. Therefore decrease in intracellular pH will reduce the Mg2+ buffer capacity of ATP [Mulquiney and Kuchel, 1997].
3.2. Effects of Diffusion
The enhanced response of Mg/Ca ratio to temperature in low-Mg foraminifera may also result from the biological organization of the calcification site. The privileged space of calcification is isolated from seawater by a membranous diffusion barrier [Angell, 1967; Hemleben et al., 1986]. If we assume that this space is kept with a low Mg concentration, it is possible that the leakage of Mg2+ from ambient seawater would increase with temperature due to the effect of temperature on the diffusion constant; and this, in turn, would amplify the Mg/Ca ratio response.
4. Ecophysiological Significance of Mg in Foraminifera Shells
Because the Mg/Ca ratio in the shell is strongly regulated by the organism it is relevant to ask what are the potential biological-ecological roles of this ratio? A few possible answers are discussed below. In some of the scenarios, Mg in the shell plays a negative role and hence needs to be removed while in others it may play a positive role.
4.1. Biological “Fixation” on the Calcite Polymorph
Biomineralizing organisms are highly conservative with regard to the identity of the mineral polymorph that they precipitate [Weiner and Dove, 2003]. Most foraminifera precipitate calcite with the exception of the Robertindae, which precipitate aragonite. Given the high Mg/Ca ratio in present-day oceans, the kinetically favored mineral to be precipitated inorganically is aragonite [Morse and Bender, 1990; Morse et al., 1997]. It is possible that when calcitic foraminifera evolved the Mg/Ca ratio of seawater was lower [e.g., Stanley and Hardie, 1998] which could favor the formation of calcite shells. Once the calcite polymorph is “fixed” genetically, foraminifera apply various strategies (discussed above) in order to conserve the polymorph despite the later increase in oceanic Mg/Ca ratios.
4.2. Removal of Mg From Calcite Decreases Its Solubility
Removal of Mg may be related to solubility considerations. It is possible that in a less supersaturated ocean than today, such as during the Eocene [Demicco et al., 2003], shells with a higher Mg/Ca ratio were more susceptible to dissolution [Bischoff et al., 1987] which could have been crucial for their survival. This may be especially relevant for planktonic foraminifera with their delicate thin shells and for many infaunal benthic species (both fossil and extant), which live in low pH interstitial water [Bernhard, 1986] that is highly undersaturated [Archer et al., 1989]. Decreasing the Mg content of their calcite would make their shells more stable.
Accordingly, it was suggested that the evolution of the different Mg/Ca ratios in foraminifera reflect the CaCO3 saturation state of surface oceans caused by variations in atmospheric pCO2. In high CO2 conditions with low CaCO3 saturation, the low-Mg groups evolved and prevailed, while in low CO2 conditions the high-Mg species (and the aragonitic Robertindae) appeared and predominated [Martin, 1995]. Along this line, changes in temperature and, particularly, changes in the latitudinal temperature gradients, which influence the solubility of CO2 (and consequently the CaCO3 saturation conditions), may have also played a role in the evolution of different Mg/Ca ratios in foraminifera.
4.3. Control of the Calcification Process
The biomineralization process requires control of the calcification rate, crystal shape, crystal orientation, preferred crystal faces and overall shell structure [e.g., Simkiss and Wilbur, 1989]. These functions are usually accomplished by the organic matrix. One essential component of the control mechanism is a nucleation inhibitor agent that is used to stop uncontrolled precipitation from the supersaturated parent solution. Mg, which is known to kinetically inhibit calcite growth, can also serve as an efficient inhibitor agent to control calcite precipitation [Giles et al., 1995]. By controlling the Mg content of the parent solution the organism may control the calcification process. This control mechanism may therefore influence the final Mg content of the shell.
4.4. Stabilization of Amorphous CaCO3
As discussed above, it was demonstrated in several biomineralization systems that amorphous calcium carbonate serves as a transient precursor phase for calcite precipitation [Beniash et al., 1997; Levi-Kalisman et al., 2002; Addadi et al., 2003; Weiner et al., 2003]. It was suggested that in these systems Mg is used to stabilize these amorphous phases by inhibition of spontaneous crystallization. This was demonstrated in inorganic precipitation experiments that showed that high-Mg calcite (>10%) is produced via an amorphous phase and that Mg is required for stabilizing the transient phase [Raz et al., 2000]. It was also shown that proteins that stabilize biogenic amorphous CaCO3 of sea urchin larval spicules require Mg2+ to induce the amorphous phase formation [Weiner and Dove, 2003]. If amorphous CaCO3 is involved in the calcification process of foraminifera, there is probably a role for Mg in its stabilization.
5. Epithelium Paradigm for Massive Transport of Mg2+
Assuming that the parent solution is based on seawater as appears from tracer incorporation and in vivo observation [Elderfield et al., 1996; Bentov et al., 2001; Erez, 2003], the foraminiferal cell should exert massive Mg transport to decrease the Mg/Ca ratio in the calcifying solution. This can be achieved by various regulation methods that are practiced routinely by eukaryotic cells. Given the large electrochemical gradient of Mg2+ across the membrane, the cell can extract Mg2+ from the privileged space to the cytosol by opening selective Mg2+ channels. Unlike for other cations (e.g., Ca2+), the eukaryotic cell seems to tolerate high intracellular transient levels of Mg2+ concentration. However, in order to keep Mg2+ homeostasis this increase should be temporary; and eventually the Mg2+ needs to be extruded from the cell. This can be achieved by separation between the Mg2+ influx and efflux either spatially and/or temporally. Temporal differentiation implies sequestration and buffering of excess Mg2+ during the calcification period and extrusion by active Mg2+ transport afterwards, between growth intervals. However, since the presumed fluxes are very high and the buffer capacity of the cell is limited, it is likely that the separation is mainly spatial; i.e., there is simultaneous Mg2+ absorption from the privileged space and Mg2+ extrusion to the seawater from a different zone of the cell (Figure 2).
In this respect, the epithelium paradigm seems to be an attractive analog. The epithelium specialized in mass delivery of ions and other constituents from one side of the tissue to the other in a unidirectional mode (e.g., the intestinal and renal epithelium [Schweigel et al., 1999; Dai et al., 2001]). This is accomplished by polarized cells which display different properties at their different faces, the apical side facing the lumen or external fluids and the basal side facing the internal body fluids (op. cit.). It is conceivable that foraminifera (although unicellular organisms) display similar polarized behavior, absorbing Mg at the basal side, which faces the privileged space and simultaneously transporting Mg out of the cell at the apical side facing the seawater. It is possible that both the temporal and spatial differentiations described above are combined to keep a net Mg influx from the privileged space while there is a net efflux from the cell to the ambient seawater.
6. A Model for Mg/Ca Response to Temperature in Low-Mg Foraminifera Based on Intrashell Variability
The response of the Mg/Ca ratio to temperature in low-Mg foraminifera probably results from a combination of the thermodynamic response and temperature-dependent physiological processes. Among the various physiological processes that are sensitive to temperature and can influence the Mg/Ca in the shell, the biologically induced intrashell heterogeneity should be considered. Early light microscopy followed by SEM studies demonstrated that the foraminiferal wall contains several sublayers [Reiss, 1957; Towe and Cifelli, 1967]. Recently, it was realized that the Mg/Ca ratio within the shell shows large variations by a factor of up to 2-4 [Nurnberg et al., 1996; Erez, 2003; Eggins et al., 2004]. These variations cannot be ascribed to environmental changes and hence may represent changes in the calcification process. In our laboratory, we found Mg intrashell variability which is inherent to the biomineralization process corresponds to two functional stages [Bentov and Erez, 2005]. The primary calcite layer, which delineates the shape of a new chamber, and is associated with the POM, is Mg rich (>10 mol%) while the secondary layers that cover and thicken the shell wall contain much less Mg (∼3 mol%). In vivo microscopic observations suggest that these two calcite phases were formed in different biomineralization pathways. When evaluating the final Mg content of a shell, the sublayers and their relative proportion must be taken into consideration. This proportion may change with various environmental factors and can also play an important role in determining the apparent temperature dependency of the foraminiferal Mg/Ca ratio. For example, the steep Mg/Ca response to temperature in low-Mg foraminifera compared to inorganic calcite [Nurnberg et al., 1996; Rosenthal et al., 1997] may be caused by a temperature-mediated increase in the proportion of the high-Mg primary calcite. In Figure 3, we simulated this scenario by increasing the ratio between the high–Mg and low-Mg phases as a function of temperature. For both phases we assumed the inorganic dependence of DMg on temperature according to Oomori et al. . We further assume a four fold difference in the Mg content between the phases, and an increase in the proportion of the high Mg phase by 0.2% per °C. The combined curve that was obtained, which represents the total Mg content, is in good fit to the observed Mg/Ca temperature dependence of planktonic foraminifera, a slope that is steeper than that of the inorganic calcite (Figure 3).
Foraminifera display large variability in the Mg content of different genera from the same environment, large Mg inhomogeneity within individual shells and significant deviations from equilibrium Mg/Ca ratios expected for inorganic calcite. All these phenomena suggest a strong biological control on the incorporation of Mg into foraminiferal shells. We propose that the high variability of Mg/Ca ratio between species represents different biomineralization pathways that developed during the evolution of foraminifera. The high-Mg calcite shells may precipitate from transient amorphous precursors and/or with the influence of an organic matrix. We propose that the low-Mg calcite shells precipitate from confined seawater parent solution that is depleted in Mg. The Mg ions may be removed from this solution by specific Mg channels into the cytosol and/or other cellular compartments, and later pumped out of the cell against a concentration gradient. Nevertheless, field and laboratory calibrations show that the Mg/Ca in foraminifera is a good recorder of ambient temperature and Mg/Ca ratio in seawater, suggesting a certain inorganic control and possibly a physiological response that may also be sensitive to these ambient conditions. A better understanding of the biomineralization processes in foraminifera and their associated vital effects, will allow us to use Mg/Ca ratios and other paleoceanographic proxies with higher degree of confidence.
The authors would like to thank the reviewers (E. Boyle and P. Dove) and the Associate Editor (P. Martin) for their helpful comments. We thank M. Edelman for critically reading the manuscript and his valuable assistance. Funding for this project was provided by the U.S.-Israel Science Foundation (grant 2000284) and German – Israeli Foundation (grant G-720-145.8/01). This work is part of the Ph.D. thesis of S. Bentov at the Hebrew University of Jerusalem.