The O2, pH and Ca2+ Microenvironment of Benthic Foraminifera in a High CO2 World

Ocean acidification (OA) can have adverse effects on marine calcifiers. Yet, phototrophic marine calcifiers elevate their external oxygen and pH microenvironment in daylight, through the uptake of dissolved inorganic carbon (DIC) by photosynthesis. We studied to which extent pH elevation within their microenvironments in daylight can counteract ambient seawater pH reductions, i.e. OA conditions. We measured the O2 and pH microenvironment of four photosymbiotic and two symbiont-free benthic tropical foraminiferal species at three different OA treatments (∼432, 1141 and 2151 µatm pCO2). The O2 concentration difference between the seawater and the test surface (ΔO2) was taken as a measure for the photosynthetic rate. Our results showed that O2 and pH levels were significantly higher on photosymbiotic foraminiferal surfaces in light than in dark conditions, and than on surfaces of symbiont-free foraminifera. Rates of photosynthesis at saturated light conditions did not change significantly between OA treatments (except in individuals that exhibited symbiont loss, i.e. bleaching, at elevated pCO2). The pH at the cell surface decreased during incubations at elevated pCO2, also during light incubations. Photosynthesis increased the surface pH but this increase was insufficient to compensate for ambient seawater pH decreases. We thus conclude that photosynthesis does only partly protect symbiont bearing foraminifera against OA.


Introduction
Ocean acidification has become a major threat to our world's oceans [1]. From preindustrial times until today, atmospheric carbon dioxide (pCO 2 ) concentrations increased from ,280 ppm to .390 ppm, and are predicated to rise to ,800 ppm by the end of this century under the IPCC business-as-usual emission scenario (WG 1, A2, [2]), which is likely to be exceeded [1,3]. The current rapid atmospheric CO 2 increase is mostly due to anthropogenic induced changes from increased fossil fuel combustion, deforestation and changes in land use and is now greater than at any time in the last 300 million years of Earth's history [4,5,6]. Not only is CO 2 a potent greenhouse gas in the atmosphere resulting in global warming, but about one third of the anthropogenic CO 2 increase is taken up by the oceans [1,7]. This uptake reduces pH and consequent carbonate saturation state (V) of the ocean surface waters, a process generally termed as 'ocean acidification' (OA). Phototrophic marine calcifiers (such as coccolithophores, foraminifera, calcareous algae and corals) strongly contribute to the cycling of carbon in our world's oceans, as part of the so called 'biological pumps' [8][9][10]. By changes in ocean chemistry ocean acidification poses a direct threat to most calcifying organisms and consequently the biological pumps [1,11,12].
However, the effect of bulk seawater pH is mediated through the diffusive boundary layer (DBL), which governs transport resistance between the bulk seawater and the organisms' surface.
Around phototrophic organisms (including most major calcifiers such as phytoplankton, foraminifera, corals and calcareous algae) DBLs can maintain substantial gradients of O 2 and pH to the bulk seawater, due to their high photosynthetic and respiratory activity [13][14][15][16][17][18][19][20][21]. Especially under daylight conditions, surface pH levels of phototrophic or photosymbiotic organisms can differ strongly (.0.1 pH units) from the surrounding seawater [13][14][15][16][17][18][19][20][21]. It is this surface pH and the resulting gradients within the organisms' DBL, rather than the bulk seawater pH, which determine ion-availability [17] and consequently transport kinetics between the tissues and surrounding seawater. Microenvironmental pH dynamics are therefore likely to play an important role in physiological responses to ocean acidification. Understanding O 2 and pH dynamics and variability within the DBLs under both present day and future OA conditions is therefore essential for all transport involving metabolic processes such as calcification, photosynthesis or respiration.
We hypothesize that OA induced increases of seawater DIC might enhance photosynthesis of photosymbiotic calcifiers and consequently result in increased pH levels on their surfaces in daylight. Thus, the pH DBL might form a shield around the organism protecting it from OA. We studied whether this pH elevation within their microenvironment can protect photosymbiotic calcifiers (or at least partly compensate) from the effects of ocean acidification in daylight and therefore lend additional resistance compared to non photosymbiotic calcifiers. We tested this hypothesis by measuring the O 2 and pH microenvironment of 4 photosymbiotic and 2 symbiont-free benthic tropical foraminiferal species under different ocean acidification scenarios in light and dark conditions. Benthic foraminifera represent a good group of model organisms for this study, because compared to most other calcifiers, calcification is periodic rather than continous, and periods of calcification can be detected visually. Additionally, the process of chamber formation is very sensitive to mechanical disturbances and thus unlikely to occur in short term flume measurements (see material and methods section, also [22][23], reviewed in [24]). Impacts of active calcification on pH microenvironments can thus be excluded during the measurements. In addition, both symbiont-free and photosymbiotic species were tested, allowing for the direct comparison of the effects of net photosynthesis and respiration on O 2 and pH microenvironments under equal experimental conditions.

Sampling and Culturing
Specimens of the photosymbiotic species Marginopora vertebralis, Amphistegina radiata, Heterostegina depressa, and Peneroplis sp., and the symbiont-free species Quinquelloculina sp. and Miliola sp. were hand collected from coral rubble and other substrates containing foraminiferal assemblages by SCUBA diving during a cruise in the summer months of 2010 in the Whitsunday area, central section of the Great Barrier Marine Park. All necessary permits were obtained prior to field collection from the Great Barrier Marine Park Authority (Permit  [35][36]. A detailed description of the sampling sites can be found in Uthicke et al. [25]. After collection, specimens were washed off substrates, cleaned by gentle washing and sieving and identified to species and genus level [26] under a dissecting-microscope (Leica MX16 A, Solms, Germany). Samples were kept in natural seawater (24u -26uC) under low light conditions (10 mmol photons m 22 s 21 ), until they were transported to the Australian Institute of Marine Science (AIMS) in Townsville. Prior to experiments, specimens acclimatised in indoor climatic chambers.3 weeks in natural seawater (replaced every 3 days, sediments removed) at 24u -26uC, 10 mmol photons m 22 s 21 , 12 h : 12 h diurnal cycling and fed with microalgae (Isocrysis sp.). Salinity of nearshore seawater available at the AIMS was diluted (32-34) due to high seasonal rainfall. During culturing and experimental treatments seawater salinity was therefore adjusted to 35 by the addition of sea salt (Sunray, Cheetham Salt, Melbourne, Australia). Salinities were measured using a refractometer (S/Mill-E, Atago, Tokyo, Japan).

Experimental Setup
Carbon perturbations experiments were performed by the addition of CO 2 enriched air into a semi-closed circulation system of filtered (1 mm) natural seawater. CO 2 enriched air (0.2%) was humidified via a system of Erlenmeyer flasks and bubbled into an aerated reservoir tank (30 L), connected to incubation chambers, which contained the organisms (water flow rate 0.5-1.0 cm s 21 ). Gas flow rates and thereby pCO 2 levels were regulated via mass flow controllers (accuracy 1.5%, GFC17, Aalborg, Orangeburg, NY, USA). The system was allowed to equilibrate for.48 h.
All amperometric and potentiometric microsensor measurements were conducted in a Faraday cage to minimize electrical disturbance. Before the measurements specimens were carefully transferred with a fine brush from the incubation chambers into a flow-cell (1.2 ml volume), connected to the same circulation system. Net flow rates within the flow cell were adjusted volumetrically to 0.5060.02 cm s 21 (mean 6 SD), to simulate average natural in situ flow conditions experienced by epifaunal and shallow infaunal foraminifera within the benthic boundary layer of reef environments [27]. Net horizontal flow was monitored ,3 mm above the foraminiferal surface by observing particle movements via a stereo-microscope (K400, Motic, Xiamen, China).
Illumination was provided from above via a fiber-optic guide from a halogen light source (Schott KL2500, Mainz, Germany). Light intensities were monitored with a quantum irradiance meter (LI-250A, LI-COR, Lincoln, NE, USA), combined with a light sensor for photosynthetic active radiation (PAR).

Experimental Procedure and Determination of Microenvironmental Dynamics
Using a fine brush, foraminifera were positioned horizontally in the middle of the flow cell resting on their central elevations, with the exception of Marginopora vertebralis, which exhibits a flat surface structure ( Figure 1). Microsensor tips were positioned on the calcite shell surfaces of foraminifera, using a stereo-microscope and a 3D-manual micromanipulator (MM33, Maerzhaeuser, Wetzlar, Germany). O 2 evolution within the DBL of phototrophic species was tested under varying light intensities (data not shown). A light intensity of 30 mmol photons m 22 s 21 was found saturating for all photosymbiotic species, without causing photo-inhibition in the tested low light species Amphistegina radiata and Heterostegina depressa [27] and used throughout all 'light' experiments (see also [33,34]).
To determine the t 90 value of steady-state signals of the system O 2, pH and Ca 2+ probes were positioned on the test surface of photosymbiotic individuals and recorded for ,30 min, while light levels were altered (light/dark changes). O 2 (pA) reached.90% steady-state signals ,2 min, pH values (mV) took,6 min, while Ca 2+ (mV) values did not change significantly. To ensure steadystate, light levels were applied for 10-60 min prior to measurements. Steady-state profiles were measured in step sizes of 50 mm (up to 400 mm) and 100 mm about 1500 mm upward perpendicular to the foraminiferal test, through the diffusive boundary layer into the bulk seawater ( Figure 1). Due to slow erecting of individuals by rhizopodial movements, gentle nudges with a fine brush were applied in between profiles to assure rhizopodial retraction, so that foraminifera and their extending DBLs remained in their horizontal position.  To estimate spatial microenvironmental O 2 heterogeneity across the shell surfaces, specimens of every species were fine scale profiled at 432 matm from front to back in flow direction ( Figure 1).

Determination of O 2 , H + and Ca 2+ Microenvironmental Dynamics
To account to some extent for spatial heterogeneity across shell surfaces ( Figure 1) during the profiling experiment, foraminiferal specimens (n = 2) were profiled in 2-4 locations on their calcite shell during the experiment (indicated by red crosses in Figure 1). To determine possible treatment effects on O 2 dynamics and to evaluate exact placement of microsensor tips for consecutive measurements, individuals were profiled with O 2 microsensor at 432 matm in light, prior to each treatment incubation. Profiling experiments were conducted at a pCO 2 of 432 matm (pH of 8.22; ambient), 1142 matm (pH 7.85) and 2151 matm (pH 7.60) with photosymbiotic species, and at two pCO 2 levels (432 and 2151 matm) with symbiont-free individuals (Table 1). After 24 h of incubation, microsensor measurements across the DBL of all specimens in both light (30 mmol photons m 22 s 21 ) and darkness were conducted for O 2 on day 2, pH on day 3, and Ca 2+ on day 4.

Monitoring of Treatments
Seawater was renewed for each experimental treatment and kept at a constant salinity (35) and pH according to the treatment (Table 1). Temperature, pH, DIC, total silicate and total phosphorus were monitored daily. DIC samples were filtered (0.2 mm nylon filters), stored gas tight, head-space free at 4uC and analysed within a week by flow injection analysis [35]. Samples for nutrient analyses (including total silicate and phosphorus) were filtered (0.2 mm nylon filters), immediately frozen and consequently analysed with a Bran and Luebbe AA3 segmented flow analyzer (Norderstedt, Germany) following Ryle et al. [36]. Samples for total alkalinity (TA) were taken at the end of each experiment, filtered (0.2 mm nylon filters), poisoned with HgCl 2 and kept at 4uC until being shipped to the University of Sydney, where they were analysed by open cell potentiometric titration [37], and calculated using linear Gran plots [38]. Corrections were applied based on certified reference material (A. Dickson, Scripps Institution of Oceanography, CA, USA).

Assessment of Individuals
For microsensor measurements, healthy, intact foraminiferal specimens of similar size and pigment shading were selected and liveliness confirmed in all individuals by the observation of movement. Individuals were photographed (Canon 30D, Tokio, Japan) via the dissecting microscope, before and after the experimental treatments (for complete sets, see Figure S1, S2, S3). At the end of the experiments, individuals were examined and photographed under a fluorescence microscope (Axioskop mot plus, Carl Zeiss, Goettingen, Germany) equipped with a digital camera (AxioCamMRc5, Carl Zeiss, Goettingen, Germany). Fluorescence images were obtained using a halogen lamp for incident light and DAPI (excitation, G365 nm; dichroic mirror FT395; emission LP420 nm) and FITC (excitation, BP 450-490 nm; dichroic mirror FT510; emission LP515 nm) filter sets (Carl Zeiss, Goettingen, Germany). Foraminiferal sizes (longest diameter) were measured in small individuals from microscopic Table 1. Experimental parameters (mean6SD), monitored each day of each 4-day pCO 2 incubations (n = 4) beside TA, which was sampled at the end of each experiment (n = 1).

Carbonate Chemistry Calculations
Calculations based on measurements of DIC, pH, temperature, salinity, total-phosphate and silicate (Table 1) were performed in CO 2 SYS [39], using K1 and K2 according to Millero et al. [40], with dissociations constants for H 2 SO 4 detailed in Dickson [41]. Measured and calculated levels of total alkalinity deviated ,0.2%, indicating that carbonate chemistries were in equilibrium throughout the experiments (Table 1).

Data and Statistical Analysis
Hydrogen ion (H + ) concentrations for dilute aqueous solutions were calculated from pH levels. Differences in concentrations between the bulk seawater and the surface of the shells, denoted as DO 2 , DH + and DCa 2+ , were attained from the measured profiles. Concentration differences were calculated as the lowest and highest spatial points of the profiles respectively. At very low metabolic rates and therefore increased resolution, profile noise was balanced by a line of best fit through the seawater baseline concentrations, and DBL gradients, to attain concentration differences. Since microsensor measurements of O 2 , pH and Ca 2+ were performed consecutively on different days, they did not depict true spatial replicates of one location (see also discussion 'Variability of microsensor measurements'). Measurement position differences of DO 2 , DH + and DCa 2+ within individuals ( Figure 1) were found to be non-significant. Consequently profiles (n = 2-4) were averaged for every individual for statistical analyses.
Means of DO 2 , DH + and DCa 2+ over replicate profiles per individual were tested for normality and homogeneity of variances by normality plots and Levene's tests, respectively. Since parametric assumptions were violated, complete data sets of mean DO 2 , DH + and DCa 2+ were analyzed by Kruskal-Wallis one way analysis of variance, and alpha levels Bonferroni corrected ( Table 2). Group comparisons were performed using Wilcoxon signed rank test (WSRT) for paired samples, Kruskal-Wallis one way analysis of variance and Wilcoxon rank sum tests ( = Mann Whitney U-tests) for unpaired samples, respectively. The ratios of mean DO 2 /DH + of all individuals were compared across pCO 2 treatment groups using generalized linear models (GLMs). All statistical analysis used the software R [42] or SPSS 13.0 (IBM, Armonk, NY, USA).

Individual Fitness
Both Heterostegina depressa at 2151 matm and Amphistegina radiata individuals at 1141 and 2151 matm showed visual signs of symbiont loss (i.e. bleaching) at the end of the 4 day incubations ( Figure S2, S3). In A. radiata, bleaching was accompanied by severe symbiont clumping within the cell body.   (Figure 1). In phototrophic specimens, DBL thickness was laterally enlarged where symbiont densities, and therefore photosynthetic activity, was higher than at the central part of the test. In A. radiata, H. depressa and Peneroplis sp., DBLs were also enlarged at the upstream edges. Differences of O 2 between the shell surface and the bulk seawater, denoted as DO 2 , varied across the shell and among individuals, and were generally strongly elevated in photosymbiotic, and slightly reduced in symbiont-free species. The downstream edge of M. vertebralis, in which symbionts were sparse, exhibited a slight O 2 under-saturation.

Time Replicated O 2 Dynamics within Individuals under Illumination
Within individuals, mean DO 2 at 432 matm (control measurements) remained constant, indicating the absence of confounding  factors (WSRT, V = 33, p-value = 0.677, Figure 3). Only in M. vertebralis at 432 and 2151 matm did DO 2 variability increase from prior to during the incubations (Figure 3). This confirms that the repeated placement of microelectrodes on individuals did not affect readings. Variation in DO 2 between individuals was greater for photosymbiotic than for symbiont-free species ( Figure 3) and highest for M. vertebralis (all pCO 2 treatments), A. radiata (prior to the incubation) and H. depressa (prior to and during the incubation) at 2151 matm. Under illumination mean DO 2 was significantly elevated at all pCO 2 treatments in photosymbiotic, compared to symbiont free species (4468 mM, vs. 20.00260.753 mM (mean6SE), U-test: W = 0, p = 1.90e 207 ). Beside individuals that exhibited symbiont loss, mean DO 2 of photosymbiotic species did not change significantly between elevated and control pCO 2 (WSRT: V = 11, p = 0.106). In A. radiata, which displayed severe visual signs of bleaching, DO 2 was strongly decreased at 2151 matm ( Figure S3). DO 2 of symbiont-free individuals remained usually negative, very low and similar at both pCO 2 treatments. Yet, some profiles of positive DO 2 (i.e. net photosynthesis) were measured in both Quinqueloculina (at 2151 matm) and Miliola specimens (Figure 3). Subsequent fluorescence imaging revealed chlorophyll autofluorescence of epiphytes on the shell surfaces of these symbiont-free individuals (Figure 4).

O 2 , H + and Ca 2+ Dynamics within and between Treatment Groups
Illumination significantly increased mean DO 2, and decreased mean DH + in photosymbiotic, compared to symbiont-free species at all pCO 2 and between light and dark, indicating net photosynthesis (Table 2, Figure 5, 6). Beside A. radiata specimens, which strongly bleached at the highest pCO 2 level ( Figure S3), mean DO 2 in light did not change significantly between pCO 2 treatments (Kruskal Wallis: X 2 = 1.8584, df = 2, p = 0.395). In darkness mean DO 2 was negative in all photosymbiotic species indicating respiration (21163 mM), which was enhanced in M. vertebralis and H. depressa at 1141 matm and reduced in A. radiata at increased pCO 2 ( Figure 5). Symbiont-free species showed net respiration in both light and dark (21.1760.54 mM).
In contrast to DO 2 , mean DH + was significantly affected by pCO 2 treatment, trophic level, species and symbiont-type (Table 2). Under illumination, mean DH + of all photosymbiotic species decreased with increasing pCO 2 (21.6760.35 nM at 432 matm vs. 23.5360.66 nM at 2151 matm, Figure 6), with the exception of Peneroplis individuals, where net photosynthesis was low and variable between the pCO 2 treatments ( Figure 5). In darkness at 432 matm, mean DH + (0.07060.019 nM) of all species was slightly increased indicating net respiration. Yet, all photosymbiotic species showed a negative mean DH + at elevated pCO 2 conditions in darkness (20.8860.21 nM, Figure 6). DH + of symbiont-free species was generally much lower in light (20.2060.15 nM), compared to photosymbiotic species and also slightly negative at 2151 matm at both light levels (20.4960.17 nM).
Changes in mean DCa 2+ were generally very low and exhibited high variation in space and time (39624 mM). Mean DCa 2+ did not change significantly with any of the measured factors (Table 2, Figure S4). At 2151 matm mean DCa 2+ was still not significantly different from 0 (23629 mM), indicating no net CaCO 3 dissolution or Ca 2+ uptake.

Ratios of Mean DO 2 /DH + Across pCO 2 Treatments
Mean DO 2 (i.e. netPS or respiration) and DH + were both quite variable across profiles within and across individuals (Figure 7). Yet, there was a significant linear correlation between mean DO 2 and mean DH + (R 2 $0.63, p tm ,0.0166) per individual for all photosymbiotic species, but not in symbiont-free species ( Table 3). The intercepts of the DO 2 /DH + correlations were significantly decreased at increased pCO 2 , except in H. depressa (Figure 7, Table 3). In symbiont-free species, mean DO 2 did not strongly

DO 2 , DH + and DCa 2+ Dynamics
To test whether OA induced increases of seawater DIC enhance photosynthesis of photosymbiotic foraminifera and consequently result in increased pH levels within their microenvironments, we conducted microenvironmental O 2 and pH measurements of photosymbiotic and symbiont-free foraminifera. In light, net O 2 evolution (photosynthesis) within the DBL of photosymbiotic species remained relatively unaffected by the pCO 2 treatments and surface pH was significantly increased. Yet, H + differences (DH + ) were significantly enlarged within the DBL with increasing pCO 2 . However, the H + decreases only amounted to ,27% (at 432 matm) and ,14% (at 2151 matm) of the ambient seawater H + concentration. Photosynthesis was thus was insufficient to compensate for the more than four-fold increased ambient H + concentrations between the highest and lowest pCO 2 treatment ( Table 4). Rates of net photosynthesis of marine phototrophs primarily depend on temperature, nutrients and light availability, as well as the efficiency of the individual carbonate concentration mechanisms (CCMs, [43][44][45]). Except for bleached individuals, DO 2 (i.e. net photosynthesis) was not influenced by pCO 2 in any species (Table 2, Figure 3). Since light levels were saturated and nutrient concentrations and temperature remained constant throughout each treatment, this may indicate either that the photosynthesis of photosymbiotic foraminifera was CO 2 saturated at ambient pCO 2 concentrations, or that a down-regulation of DIC uptake occurred at increased pCO 2 . This notion is in agreement with previous studies on diatoms [46,47] and Symbiodinium sp., both in culture and in hospite of corals [48,49] and foraminifera [50], displaying a down-regulation of CCMs and only slight effects of increased DIC on net O 2 evolution. Since there is no indication that the photosynthetic quotient (O 2 /CO 2, [51]) of the holobiont was altered at increased pCO 2 , DIC uptake should have been     constant. Increases of pCO 2 on the other hand, cause a decrease in the CO 2 uptake capacity of seawater (i.e. an increase of the Revelle factor, Table 1). This results in larger shifts of CO 2(aq) , thus H + concentrations in response to constant DIC production2/ consumption-rates (for an extensive discussion of this aspect of carbon chemistry see [18,52] and [53] Chapter 1.5). This will lead to stronger H + -gradients in response to constant photosynthesis/ respiration rates at elevated pCO 2 , as indicated by the results (Figure 6; see also [18]). It is supported by the linear regression analyses, displaying a significant pCO 2 treatment effect on mean DO 2 /DH + of most photosymbiotic species (Table 3), and by previous modeling results of microenvironmental dynamics around phytoplankton, showing increased microenvironmental H + variability at elevated pCO 2 [18].
The decreases of DO 2 , observed between 432 matm and the elevated pCO 2 conditions in A. radiata at 1141 and 2151 matm and H. depressa at 2151 matm (Figure 3), are most likely the cause of increased symbiont loss (i.e. bleaching) at elevated pCO 2 ( Figure   S2, S3, [54], [55]). Additionally, bleaching and spatial variability of symbionts (see 'variability of microsensor measurements') in A. radiata and H. depressa resulted in severe symbiont clumping and increased heterogeneity of DO 2 and DH + across their shells. This might have led to an overestimation of the mean H + difference (DH + ) in light, in respect to the mean O 2 difference (DO 2 ), by profiling areas of high symbiont densities with pH sensor and areas of low symbiont density with O 2 sensors ( Figure 5, 6). This might explain why decreases of the DO 2/ DH + intercepts in response to increased pCO 2 were less significant in A. radiata and slightly non significant in H. depressa, compared to all other photosymbiotic species (Figure 7, Table 3).
In dark, respiratory changes of DO 2 and DH + at 432 matm were minor ( Figure 5, 6). This is in agreement with previous microsensor measurements on foraminifera and diatoms [14,15,56], indicating that net respiratory O 2 fluxes are generally very low in these protists.
Interestingly, microenvironmental H + concentrations of all species were slightly decreased in darkness, compared to the bulk seawater at elevated pCO 2 ( Figure 6). One possible reason for this may be the dissolution of the calcite shell at elevated pCO 2 in darkness, causing a local increase in pH [53]. However, this is unlikely, due to the absence of significant Ca 2+ fluxes ( Figure S4), and since V Ca was super-saturated at even the highest pCO 2 (Table 1), indicating no net calcite dissolution. Another possibility could be the continued uptake of CO 2(aq) (.10 min) in the dark for CO 2 fixation in the calvin cycle. This would however imply that CO 2(aq) uptake and fixation of the holobiont outweighed respiratory CO 2(aq) production in darkness. A third explanation could be that foraminifera actively up-regulate their microenvironmental pH in darkness, via active proton pumping or antiporter exchange [23,57] into the cell, to compensate for increased seawater pCO 2 and to maintain pH homeostasis for vital cellular functions. A fourth explanation could be the excretion of nitrogen waste by the foraminifera in the dark in the form of NH 3 , which would elevate microenvironmental alkalinity, thus increase pH. The excretion of NH 3 is widely distributed among marine protists [58][59] and might be increased at elevated pCO 2 , due to increased energy demands and nutrient uptake.
Mean DCa 2+ over the shell surface was very low, but single profiles displayed strong gradients ( Figure S4). Calcification in foraminifera, i.e. chamber formation, is discontinuous and sensitive to mechanical disturbances [23,60,61]. Due to the experimental handling it can be excluded that individuals were calcifying or preparing for chamber formation .2 h before and after the measurements. Increased Ca 2+ uptake due to calcification was therefore not expected. The measured high variability and averaged low fluxes of DCa 2+ over the shell surface are in accordance with previous microsensor measurements on tropical (Marginopora vertebralis, Amphistegina lobifera, [15]) and temperate benthic (Ammonia sp., [23]) and planktonic (Orbulina universa, [16]) Table 3. Relationships between DO 2 and DH + per individual within each species at different pCO 2 treatments (Figure 7).  specimens. This indicates that Ca 2+ exchange varies over time and is not evenly distributed over the shell surface for most foraminifera, but very localized. As Ca 2+ is an important cellular ionic regulator and cytotoxic at increased cellular concentrations [62], its exchange via Ca 2+ channels in the protoplasmic membrane must be highly regulated. Distribution of Ca 2+ channels and Ca 2+ fluxes over the foraminiferal surface are most likely patchy. Ca 2+ gradients would therefore only affect a small percentage of the total foraminiferal surface area, which would lead to the generally low total Ca 2+ fluxes, but high variability in different profiles as observed ( Figure S4).

Characterizing the Foraminiferal Microenvironment
O 2 and pH DBL dynamics of photosymbiotic foraminifera and other photosynthetic calcifiers correlate in response to illumination changes, with pH dynamics exhibiting a temporal time lag following O 2 dynamics [13,16,17,19].
The extent to which surface O 2 and pH on the organisms' surface deviate from the bulk seawater depends on multiple factors, such as the photosynthetic activity of the organism, surrounding seawater flow, seawater H + -buffering capacity, diffusivity/permeability of CO 2 from its source -spatial configuration of symbiont and host, diffusional transport constrains (1-3D) and the 3D surface structure of the location [15,17,21,53,63]. Since carbonate chemistry remained constant throughout the treatments (Table 1), most prominent factors during the experiment influencing DBL dynamics, included diffusional transport constrains to and from the symbionts, micro-flow surface dynamics and location specific rates of net photosynthesis and respiration. This is illustrated by the spatial extent of the DBLs (Figure 1, 2). The thickness of the DO 2 DBL clearly decreases along middle ridges of individuals, where laminar flow velocity was highest [64,65] and underlying photosynthesis was lower, due to decreased symbiont density in that region, compared to lateral symbiont-rich parts (Figure 1, Figure S1). M. vertebralis specimens showed the steepest O 2 and pH gradients, without enlarged DBL thickness (i.e. net O 2 fluxes), indicating overall increased photosynthesis compared to all other species (Figure 1, 3, 5). Yet, ventral sides of M. vertebralis specimens locked tightly on to the inert surface of the flow chamber, thereby creating a one-dimensional diffusional barrier. The strong O 2 and pH microgradients of M. vertebralis can therefore not solely be attributed to increased photosynthesis but emerge as a combination of the underlying photosynthesis, flat surface structure (and thereby almost parallel horizontal emerging flow field), as well as one-dimensional diffusional resistance.

Variability of Microsensor Measurements
Measurement variability was high, but much higher between, than within individuals (Figure 3, 5, 6, 7), allowing for temporal replication of microsensor measurements. Variability was not unexpected due to the typically high spatial variability of O 2 fluxes and pH dynamics across the surface of photosynthetic organisms (Figure 1, [15,[19][20][21]65]) in combination with the high spatial resolution of the microsensor measurements (reviewed in [28]). Another source of variability is due to the fact that some foraminiferal species, including M. vertebralis and H. depressa, actively transport their symbionts within their cell bodies and individual chamberlets [66,67], resulting in higher variation of DO 2 ( Figure 3) and consequently DH + over time for a specific spot on their shell surface. Spatial variability of DO 2 and consequently DH + (and their means), measured within and among the individuals, was therefore expected. Yet, spatial heterogeneity within individuals (Figure 1) was not represented in the sampling, since measurement positions were not significantly different. Also DO 2 , measured before and during the 432 matm treatment under equal conditions (Figure 3) within the same individuals, remained relatively constant, confirming that the spatial placement of microelectrode measurements could be replicated.

Mixed Responses of Ocean Acidification Experiments
Several studies have reported contrary responses of increased pCO 2 on both photosynthesis and calcification on a variety of marine taxa [11,12,[68][69][70], even within phyla (reviewed in Doney et al. [1]). Possible causes for such variability are diverse, potentially including differences in calcifying-/carbonate concentration mechanisms and their coupling, tolerance levels, adaptation mechanisms, but also differences in the experimental designs and setups. Consequently, a comparison among ocean acidification studies, even within phyla, is difficult. Especially flow, as an important experimental parameter influencing the surface pH of organisms, has not been considered in many ocean acidification experiments. Yet, flow changes are well known to severely impact microenvironmental pH levels of photosymbiotic foraminifera (Figure 2, [15]) and other phototrophs in light [19,20,63]. The changes in surface pH are especially severe within static culture experiments, where DpH can change up to .1 unit (.5 nM of H + , Figure 2, [15,71]). Zero-flow conditions for ocean acidification studies should therefore be avoided, as they are ecologically unrealistic and also confuse the carbonate chemistry of the intended pCO 2 perturbation, causing unrealistically high/low microenvironmental pH conditions in light/dark, despite increased DIC levels.
Some of the observed variability in calcification responses of photosymbiotic foraminifera to OA are likely due to differences in calcification mechanism (also discussed in [84][85][86]), as well as solubility differences of the calcite tests [87,88] of the different groups. This is represented in the literature showing unaffected or increased calcification rates in hyaline (low Mg-calcite: less soluble) and decreased rates in miliolid (high Mg-calcite: more soluble) species in response to elevated pCO 2 [70,73,76,77]. These taxa specific differences are in line with previous studies on foraminiferal DIC uptake mechanism, showing almost linear increases in miliolid Amphisorus hemprichii and almost no change in Amphistegina lobifera in response to increasing DIC (and CO 3 22 ) concentrations in the OA range between 2 and 3 mM ( [84], also discussed in [76]). Additionally, these ideas are supported by recent field studies investigating foraminiferal assemblages at volcanic CO 2 vents in the Mediterranean [79,81] and in tropical coral reefs [80]. The studies in the Mediterranean reported significantly reduced numbers of calcareous species, a complete absence of miliolid and only the presence of hyaline species at elevated pCO 2 [79,81]. The study investigating cold CO 2 seeps within tropical coral reefs reported almost complete absence of the larger epibiotic miliolid species Marginopora vertebralis and reduced species richness and diversity of sedimentary foraminifera at high pCO 2 sites [80]. A very recent study investigated symbiont-free hyaline foraminiferal assemblages in a CO 2 enriched, benthic habitat in the southwestern Baltic Sea [78]. This study showed that mainly sediment V Ca under-saturation, rather than the pCO 2 levels of the sediments, determines the population density of the benthic shallow infaunal species Ammonia ammoriensis, yet not of Elphidium incertum [78]. These findings support the idea of increased resistance/adaptation of hyaline species within their natural habitat to high pCO 2 conditions, compared to miliolid species.
The findings of this study indicate that photosynthesis can only to a minor extend compensate for ambient seawater pH decreases within the microenvironment of photosymbiotic foraminifera (Table 4). Symbiont-free and photosymbiotic foraminifera are thus likely to experience strongly decreased microenvironmental pH conditions at future pCO 2 , making their cell bodies susceptible to the physiological effects of ocean acidification.