Role of Extracellular Carbonic Anhydrase in Dissolved Inorganic Carbon Uptake in Alkaliphilic Phototrophic Biofilm

Alkaline Soda Lakes are extremely productive ecosystems, due to their high dissolved inorganic carbon (DIC) concentrations. Here, we studied the dynamics of the carbonate system, in particular, the role of extracellular carbonic anhydrase (eCA) of an alkaliphilic phototrophic biofilm composed of bacteria enriched from soda lake benthic mats. By using measurements with microsensors and membrane inlet mass spectrometry, combined with mathematical modeling, we show how eCA controls DIC uptake. In our experiments, the activity of eCA varied four-fold, and was controlled by the bicarbonate concentration during growth: a higher bicarbonate concentration led to lower eCA activity. Inhibition of eCA decreased both the net and the gross photosynthetic productivities of the investigated biofilms. After eCA inhibition, the efflux of carbon dioxide (CO2) from the biofilms increased two- to four-fold. This could be explained by the conversion of CO2, leaking from cyanobacterial cells, by eCA, to bicarbonate. Bicarbonate is then taken up again by the cyanobacteria. In suspensions, eCA reduced the CO2 leakage to the bulk medium from 90 to 50%. In biofilms cultivated at low bicarbonate concentration (~0.13 mM), the oxygen production was reduced by a similar ratio upon eCA inhibition. The role of eCA in intact biofilms was much less significant compared to biomass suspensions, as CO2 loss to the medium is reduced due to mass transfer resistance.


INTRODUCTION
Alkaline Soda lakes are extremely productive due to the high DIC availability, and can be found in various geographical locations around the globe (Melack, 1981;Priscu et al., 1982;Kompantseva et al., 2009). Soda lakes are believed to have existed throughout the geological record of Earth, and are abundant in dry terrestrial biomes. These lakes support the growth of an large array of microorganisms that are of ecological and economic importance (Antony et al., 2013). Microalgae that thrive in the high pH and salinity can highly efficiently photosynthesize due to the elevated dissolved organic carbon (DIC) levels. This makes biofilms from alkaline environments potentially useful for carbon capture. The high medium pH can scrub effectively CO 2 from the exhaust gasses from large CO 2 producing industrial units, elevating the DIC in the medium to very high levels, while phototrophs can maintain the high pH in the scrubber liquid. Initial studies on the phototrophic microorganisms of natural alkaline and saline lakes, especially the biofilm forming cyanobacteria, demonstrated their very high phototrophic activity (Sharp et al., 2017). A mechanistic understanding of the DIC uptake in the biomass is desired and the aim of this study.
The interconversion between CO 2 and H 2 CO 3 is relatively slow, with an 90% equilibration in 20 s under typical conditions, while the rest of the carbonate system equilibrates instantly. The highly effective enzyme carbonic anhydrase (CA) accelerates the reversible hydration of CO 2 . For convenience, the species CO 2 and H 2 CO 3 are often taken together and the hydration of CO 2 is then noted to produce bicarbonate (HCO − 3 ) (Giordano et al., 2005).
Phototrophic organisms are able to take up both CO 2 and HCO − 3 as the carbon source for photosynthesis, whereby an essential difference is that CO 2 can passively pass membranes and HCO − 3 is actively taken up by transporters, driven by membrane potentials or ATP. Thus, uptake of HCO − 3 can be controlled by the cell through adjusting the amount and affinity of the transporters. In photosynthesis, during carbon fixation, Ribulose-1,5-bisphosphate carboxylase/oxygenase (RuBisCo) only accepts CO 2 as the substrate, and HCO − 3 is first converted to CO 2 (Giordano et al., 2005).
We aimed to investigate how eCA activity impacts DIC uptake and productivity. With organisms enriched from sodalake benthic mats, an alkaliphilic phototrophic biofilm was created as a model system. The DIC uptake and the role of eCA in the DIC dynamics of this model system were investigated with a combination of methods, including microsensors, membrane inlet mass spectrometer, and mathematically modeling. We tested the hypothesis that the eCA is important for DIC uptake, especially at low bicarbonate concentrations. We expected that the cells will increase eCA activity when the DIC is limiting photosynthesis.

Sample Preparation
The biomass used for inoculating the investigated biofilm was prepared as described previously (Sharp et al., 2017). In short, phototrophic benthic mat samples were collected from four soda-lakes on the Cariboo Plateau, British Columbia in May 2015. Mat samples were homogenized and mixed in equal wet weight proportions. Phototrophs were enriched in flat panel photobioreactors (PBRs) with high pH and high alkalinity medium ("3e medium, " Table 1). The PBRs were exposed to 80 µmol photon m −2 s −1 red light (fluorescent lamp equipped with a red-pass filter, Congo Red) under a 16:8 light/dark cycle at 25 ± 2 • C. During cultivation, the medium was exchanged regularly and the biomass in the PBR was harvested once every week. The enriched biomass was dominated by the cyanobacterium Phormidium kuetzingianum (>50% relative abundance) (Sharp et al., 2017).
Biomass was harvested from the photobioreactors, resuspended in medium, allowed to settle for 1 h, after which the supernatant was replaced with DIC free medium and resuspended. After 5 min, the biomass was concentrated by centrifugation (500 × G, 1 min), and the pellet resuspended in DIC free medium. This process was repeated twice. Agar plates (4% and amended with corresponding media) were submerged in "3e medium" modified to have 0, 0.01, 01, 0.5, or 1 M DIC. Then the concentrated biomass was inoculated onto agar plates. Biofilm formation on the agar plate occurred within 12 h. The DIC concentration was adjusted by decreasing the NaHCO 3 concentration of the "3e medium, " compensated by NaCl to keep Na + constant. Due to equilibration with air, 0 M DIC actually contained about 0.13 mM DIC. Solidified agar plates each with a surface area of 23 cm 2 were used as substrate for biofilm support. For every DIC concentration, 2 such agar plates were inoculated, and cultivated in their respective media using halogen lamps (General Electric, USA) with continuous illumination and a light intensity of 100 µmole·m −2 ·s −1 at 23 • C, submerged in a flow cell (2 cm under surface). Each flow cell * system had a total volume of 1.5 L, including the recirculation reservoir. The agar plates were cultivated for 5-7 days without changing the medium before measurement, the medium was circulated using a peristatic pump with a flow rate of 0.65 mL·min −1 .

Microsensor Measurements
Profiles of oxygen (O 2 ) concentration (Revsbech, 1989) and pH (Jensen et al., 1993) were measured using microsensors with tip diameter of maximally 15 µm. The LIX pH sensors were prepared with protective protein coating (De Beer et al., 1997). Gross photosynthetic productivity was measured by the light/dark shift method using an oxygen microsensor with <0.2 s response time (Revsbech et al., 1981;Revsbech and Jørgensen, 1983). The microsensors were prepared, calibrated and used as described previously (Revsbech et al., 1981;Revsbech and Jørgensen, 1983;Revsbech, 1989;Jensen et al., 1993;Epping et al., 1999). The agar plate with the biofilm was place in 300 mL of freshly prepared medium (with identical composition as during cultivation). The water column above the biofilm surface was 2 cm. The sensors were positioned perpendicularly to the biofilm surface. The sample was equilibrated for 10 min before the first microsensor measurement. The medium was mixed by an airstream across the medium surface. The measurements were first performed on biofilms without inhibitor. Then, an eCA inhibitor, acetazolamide (AZ) (Mercado et al., 2009;Hopkinson et al., 2013) was injected 3-4 min before the measurement into the measurement chamber (final concentration of 150 µM). During microsensor measurements, the illumination was identical to that applied during biofilm cultivation. Calibration of the oxygen sensors was performed using N 2 and air bubbled "3e medium, " the O 2 concentration of air saturated medium was calculated using the ambient temperature and salinity of the respective media (Sherwood et al., 1991). Calibration of the pH sensors was performed using commercial buffer solutions (pH 7 and 9, Fluka, Germany).

Membrane Inlet Mass Spectrometry (MIMS)
MIMS measurements were carried out using an Agilent 5977A MSD (Agilent, USA) mass spectrometer connected to a custommade membrane inlet cuvette, the construction of the cuvette has been described previously (Rost et al., 2007;Beckmann et al., 2009). A PTFE membrane (10 µm thick, 0.33 cm 2 surface area; Reichelt Chemical Technique, Germany) was selected as the inlet. The 6 mL cuvette had an injection port at the top of the cuvette. A suspended magnetic stirrer was positioned directly above the membrane, and the temperature of the cuvette was kept constant with a build-in water jacket at 20 • C. Biomass was harvested from the agar plates, washed twice in DIC free medium, and consequently concentrated by centrifugation (1,000 × G, 30 s), then homogenized by forcing the biomass through a 0.2 mm inner diameter syringe needle several times. This procedure broke the biofilm into small and relatively uniformly sized pieces (ca. 0.1 mm in diameter). Microscopic inspection of the concentrated biomass and the absence of pigments in the supernatant showed that the procedure does not disrupt cells. To determine the ash free dry weight (AFDW) of the biomass, the biomass was first washed twice with DIC free medium, then dried at 70 • C overnight. The AFDW was calculated as the weight loss after combustion at 550 • C for 3 h. The eCA activity was measured with a method as described previously (Palmqvist et al., 1994). The MIMS cuvette was covered to exclude all light. DIC solution, prepared using NaH 13 C 18 O 3 , was injected into the MIMS cuvette containing DIC free medium buffered to pH 8.7 (Tris buffer) to achieve a final DIC concentration of 1 mM. The log enrichment of 13 C 18 O 2 was then monitored for at least 10 min. The log-enrichment of 13 C 18 O 2 was calculated as: (2) IC with the different subscripts are the mass spectrometer counts of the different 13 CO 2 species.
Subsequently, 0.6 mL of the processed biomass was injected into the cuvette suspension (corresponding to a final biomass AFDW concentration between 0.18 and 0.26 mg·mL −1 in the cuvette), and the change of the log-enrichment of 13 C 18 O 2 was monitored for a further 10 min. Due to its catalytic activity, eCA will increase the 18 O lost rate of 13 C 18 O 2 , thus accelerate the decrease of the log enrichment of 13 C 18 O 2 . Consequently, eCA activity can be quantified as the change in the decay rate of the log-enrichment of 13 C 18 O 2 per g AFDW (Palmqvist et al., 1994). An example of the acquired time vs. log enrichment chart is given in Figure 1, and the eCA activity was quantified as: in Equation (3), S 1 and S 2 are the slope of the log-enrichment of 13 C 18 O 2 change before and after injection of biomass, respectively, as shown in Figure 1, and AFDW is the ash-free dry weight of the biomass injected in this particular measurement.
The MIMS setup was also applied to investigate DIC uptake of the biomass harvested from the agar plates: The MIMS cuvette was filled with "3e medium" modified with 0.1 M DI 13 C prepared using NaH 13 CO 3 and buffered to pH 8.7 using Tris buffer. To initiate the measurement, DIC free medium washed and homogenized biomass was injected in the dark into the cuvette (final AFDW concentration between 0.18 and 0.26 mg·mL −1 ). After 3 min, 100 µmole·m −2 ·s −1 light was supplied by turning on a halogen lamp. The signals of O 2 , 12 CO 2 , and 13 CO 2 (m/z value 32, 44, 45, respectively) were monitored. After the steady state was reached (indicated by linear behavior of the signals, achieved within 5 min), the signals were monitored for at least 5 more minutes to acquire steady state rates (linear phase slope). Then, an eCA inhibitor, dextran-bound-acetazolamide (DBAZ) was added into the cuvette to a final concentration of 150 µM (Hopkinson et al., 2013). As DBAZ cannot enter the cell, it inhibits only eCA. The aforementioned signals were monitored further for at least 10 min, and the linear slopes were calculated. Finally, to acquire dark respiration rates, the light was turned off for the last 10 min of the measurement. To calculate the absolute rate, the MIMS signals were calibrated using linear calibration curves (for O 2 , N 2 and air bubbled "e3 medium"; for 13 CO 2 , N 2 bubbled DIC free "3e medium, " 0.001 M and 0.1 M DI 13 C "e3 medium, " buffered to pH 8.7). Also, the consumption of O 2 by the MIMS, and the effect of temperature and salinity on O 2 concertation were corrected (Sherwood et al., 1991). The equilibrium constants, K 1 and K 2 , of the carbonate system for calculating various DIC concentrations was given in Table 3 (Davies, 1962;Eigen, 1964;Steiner et al., 1975;Johnson, 1982;DOE, 1994;Millero et al., 2002;Schulz et al., 2006;Wolf et al., 2007;Lee and Rasaiah, 2013). If not otherwise specified, all curves in the figures were created using one phase log-association provided by the software Prism (Graphpad, USA).

Mathematical Model
A mathematical model was constructed to analyze the observed trends in the MIMS measurements. The model was based on the concept of Schulz et al. (2006), modified to incorporate biological processes (Figure 2). In short, the dynamic model calculates fluxes in and out of cells, based on mass balance equations, taking into account diffusion, chemical processes (e.g., acidbase reaction), biological processes (e.g., CO 2 fixation). As we considered only the fate of added 13 CO 2 label, dark respiration was not included. The equations describing the considered processes are given in Table 2. The values of the used parameters are given in Table 3. The biological processes were considered to occur only in the cells, and eCA was considered to be cell-bound only. The DIC dynamics are calculated in the bulk liquid, i.e., the free-flowing medium in the MIMS measurement chamber, using the biomass conversion rates and fluxes between biomass and medium. The boundaries were set to be the cell surface and the interface between the flow boundary layer and the bulk liquid, respectively. Initial concentrations of the dissolved species were set to the same values as in air saturated, freshly prepared "3e medium" with 0.1 M DI 13 C, buffered to pH 8.7 with 20 mM Tris. The effect of eCA was modeled as pH dependent enhancement factors, F eCA , for reaction rates of CO 2 hydration and HCO − 3 dehydration (i.e., multiplied to the reaction rate constant) (Coleman, 2000;Supuran, 2016). This pH dependent enhancement factor is controlled by an imposed pH independent factor, F eCA . Higher F eCA values represent stronger eCA activity, F eCA = 1 indicates no eCA activity. eCA inhibition was simulated For clarity, 13 C was simplified as C, and H + and OH − ions are omitted. For more detailed description of the processes and parameters illustrated here, refer to Tables 2, 3.
Frontiers in Microbiology | www.frontiersin.org as reducing the reaction rate of CO 2 +H 2 O ⇋ HCO − 3 + H + to un-catalyzed values (i.e., set F eCA to 1, refer to Tables 2, 3 and Figure 2). The model, which was a dynamic model represented by a combination of partial differential equations (Table 3), was solved by first converting the partial differential equations to a system of ordinary differential equations (ODEs), using the method of lines (Wouwer et al., 2014). The ODEs were then solved with the ode15s solver provided in the MATLAB software (version 2015b, MathWorks, USA). A parameter analysis of the model was performed by varying the values of model different parameters (refer to Table 3 and Figure 2).

Profiles of Dissolved Oxygen, pH, and Gross Photosynthetic Productivity
O 2 production was observed in all samples under illumination (Figure 3). The thickness of the biofilm increased with higher DIC concentrations in the culture medium (visual observation) and also the O 2 production increased with increasing DIC concentrations. A maximum O 2 concentration of 2.3 mM was measured in biofilms cultivated with 1 M DIC. The pH increased from 8.7 at the surface to 10 in biofilms cultivated with low DIC (∼13 mM) when eCA inhibitor was not present. In biofilms cultivated in media with high DIC (>0.5 M DIC), the pH increased less or not at all due to increased buffering of the carbonate system. After eCA inhibition, oxygen production decreased, especially at low DIC (oxygen flux was reduced from 1.7 to 1.3 ·µmole·m −2 ·s −1 ). The pH decreased only at low DIC, due to the low buffering strength of the medium. Biofilms cultivated at higher DIC concentrations suffered less productivity loss after eCA inhibition.

Extracellular Carbonic Anhydrase and Dissolved Inorganic Carbon Uptake of Resuspended Biofilms
All samples exhibited eCA activity, which decreased with increasing DIC concentration during cultivation. Above a DIC addition of 0.5 M no further decrease was observed. The effect of DIC addition on eCA activity could be four-fold (Figure 4). The dynamics of O 2 , 13 CO 2 , and 12 CO 2 in response to illumination and eCA inhibition were similar for biomass grown at different DIC concentrations (Figure 5). Upon illumination, O 2 started to increase, and, remarkably, 13 CO 2 also increased, indicating CO 2 was released from the cells. Addition of DBAZ (an eCA inhibitor) induced a sudden release of 13 CO 2 , after ∼2.5 min, 13 CO 2 increased steadily again. The sudden release of 13 CO 2 is likely the combined result of unbinding of eCA *TABLE 2 | State variables and their corresponding expressions in the proposed model.

State variable
Diffusion **Chemical reaction ***Biological process Table 3 and Figure 2; ** F eCA equals 1 except at the surface of the cell and when no DBAZ is present; equals 10 12 for compartment inside the cell to simulate the effect of iCA. *** Only occurs in biomass (i.e., modeled as sink/source at the surface of the biomass clump).
Frontiers in Microbiology | www.frontiersin.org *    bound 13 CO 2 due to addition of DBAZ, and the shift in equilibrium of the carbonate system caused by eCA inhibition (Palmqvist et al., 1994). Whereas, the steady increase is the leakage of 13 CO 2 into the bulk medium. DBAZ addition caused a decrease of O 2 release for high DIC adapted biofilms (0.1, 0.5, and 1 M DIC) but, remarkably, accelerated O 2 production in low DIC adapted biofilms. The increase in 13 CO 2 was steeper than when eCA was not inhibited, indicating an enhanced 13 CO 2 release and/or a slower 13 CO 2 depletion rate, e.g., conversion to bicarbonate, after eCA inhibition. After the light was turned off, O 2 decreased immediately. 13 CO 2 continued to increase for a short period of time, after which the release of 13 CO 2 stopped. From data in Figure 5, oxygen production rates, 13 CO 2 release rates, and apparent gross DIC uptake rates (calculated as the sum of net oxygen production and 13 CO 2 release but does not account for the conversion of released 13 CO 2 into bicarbonate) were calculated (Figure 6). Without eCA inhibitor, net O 2 production and 13 CO 2 release rates increased with supplemented DIC concentration during cultivation and leveled off at about 0.1 M DIC. The apparent gross DI 13 C uptake rates increased after eCA inhibition. Before inhibition of eCA, around 50% of the apparent gross DI 13 C uptake was released as 13 CO 2 , whereas when eCA was inhibited, up to 87% of the apparent gross DI 13 C uptake was released as 13 CO 2 .

Modeling of DIC Dynamics
Similar to observed experimentally, the model suggested that, under certain conditions, net O 2 production rate can be increased by inhibiting eCA (Figure 7). At lower K S,HCO − 3 values (<1 × 10 −2 M·s −1 , thus a high affinity to HCO − 3 ) net oxygen production rate increased after the simulated eCA inhibition ( Figure 7A). When other parameters were kept constant, and K S,HCO − FIGURE 3 | Profiles of gross photosynthetic activity (Gp), oxygen concentration, and pH measured with microsensors in the investigated biofilm. Row (A-C) give the profiles acquired from biofilms cultivated with medium supplemented with 0, 0.5, and 1 M dissolved inorganic carbon (DIC), respectively. Y-axes give the depth relative to the surface of the measured biofilm, 0 indicates the approximated position of the surface of the biofilm. Black bars represent Gp after AZ addition, empty bars show the difference of Gp before and after AZ addition. Solid lines represent profiles (i.e., O 2 , pH) before AZ addition, dashed lines represent profiles (i.e., O 2 , pH) after AZ addition. The data shown here are the mean values of 3 replicate measurements, for gross productivity, black and white error bars present the standard deviation for triplicates before and after AZ addition, respectively. Error bars for oxygen and pH measurements are omitted for clarity.
inhibitor of eCA, acetazolamide (AZ) was applied to the biofilm during the microsensor measurement. Inhibition of eCA led to decreases in both O 2 concentrations in the biofilm and gross photosynthetic productivities, especially for low DIC (∼13 mM) cultivated biofilms. This shows directly the presence of eCA and its positive effect on photosynthetic oxygen production. The activity of eCA was further verified by the MIMS measurements (Figure 4). Previous studies investigating the role of eCA in DIC uptake mainly focused on planktonic microalgae, exposed to much lower DIC concentrations and lower pH values than in the present study (Palmqvist et al., 1994;Nimer et al., 1999;Rost et al., 2007;Beckmann et al., 2009;Kupriyanova et al., 2011;Hamizah et al., 2017). Our results confirmed the activity of eCA and showed its participation in bicarbonate uptake by biofilm-inhabiting cyanobacteria.
The fact that the activity of eCA is downregulated at increasing DIC levels, further demonstrates its role in DIC uptake. The function of eCA is to accelerate the extracellular interconversion from CO 2 to bicarbonate, i.e., to a DIC species that can be actively taken up by the cells. As a result, DIC uptake is enhanced and subjected to regulation by the cells (Giordano et al., 2005). Most importantly, eCA prevents leaked out CO 2 from escaping from periplasmic space into the bulk liquid. Most of the CO 2 reaching the periplasm originates from inside the cells, where it is produced from HCO − 3 . The microsensor measurements showed that, in biofilm cultivated with no additional DIC, the pH decreased after eCA inhibition (Figure 3), indicating a lower rate of CO 2 fixation.
Both CO 2 and HCO − 3 can serve as extracellular carbon sources, whereby CO 2 is taken up passively by diffusion and bicarbonate by active transport. CO 2 uptake does not cost energy (i.e., passive diffusion), but cannot drive accumulation of DIC in the cell (Giordano et al., 2005). Active HCO − 3 uptake requires energy supplied by photosynthesis (Giordano et al., 2005). Cyanobacteria have 3 known active HCO − 3 uptake pathways for accumulation of DIC inside their cells (Omata et al., 1999;Shibata et al., 2002;Price et al., 2004Price et al., , 2008Raven et al., 2008). Under the experimental conditions in this study (i.e., pH buffered to 8.7) the intracellular DIC level will be higher than that in the medium, as cyanobacteria cells can concentrate DIC intracellularly up to 1,000-fold the extracellular level. If in the cytoplasm the pH is 7.3, the CO 2 concentration in the cytoplasm is likely higher than that in the bulk medium as a result of a much higher total DIC concentration and a lower pH, even if the carbonate system is not in equilibrium. In the cyanobacterial cells, HCO − 3 is transported to the carboxysomes, and converted to CO 2 for RuBisCo consumption, thus the CO 2 concentration inside the carboxysomes can be further elevated (Price et al., 2008;Raven et al., 2008;Kerfeld and Melnicki, 2016). These CO 2 concentration gradients lead to the observed 13 CO 2 release under illumination during the MIMS measurement.
In suspension (i.e., during MIMS measurement) inhibition of eCA leads initially to a steep increase in 13 CO 2 release. After this initial increase, increases in the steady state 13 CO 2 release rates was also observed (Figure 6). This shows that eCA reduces the CO 2 leakage into the bulk medium. Interestingly, the apparent gross DIC uptake rates were increased markedly after eCA inhibition in the experimental MIMS measurement (Figure 6C). A similar trend was also predicted by the proposed FIGURE 6 | Net O 2 production rate, 13 CO 2 release rate, and apparent gross dissolved inorganic carbon (DIC) uptake rate (calculated as the sum of Net O 2 production rate and 13 CO 2 release rate) before/after DBAZ injection of the membrane inlet mass spectrometry (MIMS) DIC uptake experiments. X-axis gives the supplemented DIC concentrations applied during cultivation. Y-axes in (A-C) give the calculated absolute rates of net O 2 production, 13 CO 2 release, and apparent gross DIC uptake rate, respectively (square and circle indicate rates before and after DBAZ injection). model (Figure 7). This suggests that the enhanced CO 2 leakage could be compensated by a higher HCO − 3 uptake, either by promoting active HCO − 3 uptake or by increasing HCO − 3 availability. Figure 6 shows, when eCA was active, 50% of the DIC uptake was assimilated or converted to bicarbonate for reuptake, the remainder escaped by CO 2 leakage into the bulk medium. When eCA was inhibited, for higher DIC cultivated biomass, only <13% of the DIC uptake could be assimilated or converted to bicarbonate, while the rest escapes into the bulk medium.
The positive effect of eCA on DIC uptake could also be observed in the biofilm, especially for the biofilm cultivated at low DIC. For high DIC biofilms, the effect was much less pronounced. Mass transfer in biofilms is limited by diffusion, leading to significant mass transfer resistance. Such a resistance does not exist in suspensions (Dibdin, 1997). Consequently, CO 2 that has leaked out of cells, is largely retained in the gives the net O 2 production rate, (D) shows 13 CO 2 release rate. In all panels, the solid lines give the rates before eCA inhibition; the dashed lines give the rates after eCA inhibition.
biofilms and not lost into the medium. It thus remains available for re-uptake. In this scenario, where generated intermediates constantly leak out of cells, mass transfer resistance can enhance the uptake of substrate. At low DIC concentration (∼0.13 mM DIC), DIC can be limiting inside the biofilm, making the eCA more important. At higher DIC concentrations, DIC becomes more readily available inside the biofilm, thus the eCA activity was less important (Figure 4).
Thus, the function of eCA is probably the promotion of DIC assimilation (i.e., photosynthetic carbon fixation efficiency). Remarkably, in low DIC cultivated biomass, inhibition of eCA induced a stimulation of the photosynthesis rate in suspensions ( Figure 6A). To better understand this observation, an analysis of the system was performed using the dynamic model. Microalgae adapt to low bicarbonate concentrations by increasing their affinity for bicarbonate (i.e., decrease the half saturation concentrations for bicarbonate uptake), that also reduces the maximum uptake-and assimilation rate of bicarbonate (Aizawa and Miyachi, 1986;Raven et al., 2008). A biofilm that is adapted to low bicarbonate concentrations might exhibit only a slight increase in photosynthetic rate at high bicarbonate concentrations, as it is limited by lower maximum rates. Conversely, a biofilm adapted to high bicarbonate concentrations will become bicarbonate limited at low bicarbonate concentration.
With this concept in mind, the remarkable increase in O 2 production after eCA inhibition for low DIC cultivated biomass ( Figure 6A) can be explained. After transferring the biomass to a higher DIC concentration, the bicarbonate uptake is limited by the maximum uptake rate (e.g., limited by the number of bicarbonate transporters). As eCA converts CO 2 to bicarbonate, its inhibition increased the concentration of dissolved free 13 CO 2 in the cell boundary (e.g., the periplasm) that always can pass the membranes. Since this biomass has low maximum uptake rate for HCO − 3 , the HCO − 3 uptake is over-saturated when the biomass was transferred to higher DIC medium (13 mM during cultivation and 0.1 M DIC during the MIMS measurement). In this special case eCA inhibition increased the CO 2 concentration in the biomass without affecting HCO − 3 uptake, leading to higher O 2 production rate. Too further clarify, the MIMS measurements were performed in comparatively much higher DIC concentration for biomass cultivated at low DICs. Our hypothesis is that this biomass has adapted to the low DIC environment, i.e., the HCO − 3 transport of these biomass have high affinity to HCO − 3 but low maximum HCO − 3 uptake rate. This means during the MIMS measurement, a decrease in periplasmic HCO − 3 concentration can still support the same HCO − 3 uptake rate as before eCA inhibition, as the HCO − 3 concentration is still high enough for low DIC adapted biomass to maintain close to maximum uptake rate (as before eCA inhibition). However, a higher CO 2 concentration due to a slower conversion rate to HCO − 3 in the periplasmic space increased CO 2 concentration in the periplasmic space, and thus slowed the leakage of CO 2 , making CO 2 more available inside the cells for assimilation. For high DIC adapted biomass, at 0.1 M DIC, the DIC uptake is limited by the lower concentration rather than the maximum uptake rate, i.e., under-saturated. Inhibition of eCA decrease the availability of bicarbonate, thus causing the net productivity to decrease. The observed increase of photosynthesis upon inhibiting eCA, in this special case, was indeed an outcome of the model (Figure 7).

CONCLUSION
To summarize, eCA converts CO 2 escaping from the cytoplasm into the periplasmic space into bicarbonate, which can be taken up again by the cell. In suspensions, eCA reduced the CO 2 leakage to the bulk medium from 90 to 50%. In biofilms cultivated with low DIC, the oxygen production was reduced by more than 25% upon eCA inhibition. The role of eCA in biofilms was much less significant at high DIC (0.5-1 M). Despite a stronger eCA activity, lower DIC adapted biomass exhibited lower net productivity and lower apparent gross DIC uptake rate. Both eCA production and bicarbonate uptake consumes energy, with the former dependent on the amount of eCA produced/maintained, and the latter dependent on the amount of uptake and the concentration of bicarbonate. Consequently, it can be suspected: the biofilm adapts to high DIC concentrations by decreasing the activity of eCA and increasing the DIC uptake rate.
To further verify this hypothesis, future studies on the DIC uptake kinetic (e.g., MIMS DIC uptake measurement with a range of DIC concentrations) should be performed. In the present study, all MIMS measurements were performed using biomass homogenized across the depth of the entire biofilm. It is possible, due to gradients (e.g., pH, light) in the investigated biofilms, biomass at different depths of the biofilm have different eCA activities and/or DIC uptake related parameters. Thus, similar measurements on biomass acquired from different depths of the biofilm (e.g., by sectioning the biofilm) should also be carried out in future studies.

AUTHOR CONTRIBUTIONS
TL performed most of the experimental, theoretical work, and the writing. CS and MA performed part of the experiment. All authors took part in the discussion of the data and participated in the writing of the manuscript.