Enhanced sequestration of carbon dioxide into calcium carbonate using pressure and a carbonic anhydrase from alkaliphilic Coleofasciculus chthonoplastes

Abstract CO2 in the atmosphere is a major contributor to global warming but at the same time it has the potential to be a carbon source for advanced biomanufacturing. To utilize CO2, carbonic anhydrase has been identified as a key enzyme. Furthermore, attempts have been made to accelerate the sequestration via pressure. This study aims to combine both approaches to achieve high sequestration rates. The carbonic anhydrase of the alkaliphilic cyanobacterium Coleofasciculus chthonoplastes (cahB1) and bovine carbonic anhydrase (BCA) are introduced into a high‐pressure reactor to catalyze the hydration of CO2 at up to 20 bar. The reactor is filled with a CaCl2 solution. Due to the presence of Ca2+, the hydrated CO2 precipitates as CaCO3. The impact of the carbonic anhydrase is clearly visible at all pressures tested. At ambient pressure a CO2 sequestration rate of 243.68 kgCaCO3/m3 h for cahB1 was achieved compared to 150.41 kgCaCO3/m3 h without enzymes. At 20 bar the rates were 2682.88 and 2267.88 kgCaCO3/m3 h, respectively. The study shows the benefit of a combined CO2 sequestration process. To examinate the influence of the enzymes on the product formation, the precipitated CaCO3 was analyzed regarding the crystalline phase and morphology. An interchange of the crystalline phase from vaterite to calcite was observed and discussed.


INTRODUCTION
Climate change is identified as a major challenge of contemporary society. Driven by greenhouse gases in general, carbon dioxide (CO 2 ) as the most common one is a principal contributor to global warming. Since the beginning of industrialization, human activities are responsible for an ongoing increase of CO 2 in the atmosphere [1]. Considering the rapid growth of world's population and a continuous industrialization process in emerging industrial countries, the planet is strongly threatened by climate change and subsequently by natural disasters, increased sea level and temperature [2,3]. In 2020, atmospheric CO 2 reached consistently over 410 ppm which represents a nearly 50% increase compared to its pre-industrial level (https://www.esrl.noaa.gov/gmd/ccgg/trends/global. html, November 2020).
On the other hand, this reservoir is a cheap and abundant carbon source, which is needed for a sustainable industrial biotechnology [4]. Current research focuses on the role of CO 2 as a substrate either in a fermentative whole cell approach [5] or in an enzymatic system [6]. Both attempts aim to produce fuels or platform chemicals as valuable products. Therefore, not only the capturing but an efficient CO 2 sequestration as a bioavailable C1 compound is a key challenge. When it comes to biological CO 2 sequestration, carbonic anhydrase (CA) is proven to be a central enzyme [7]. Promising studies were made with CA in an active membrane which is able to separate CO 2 out of a gas stream with high selectivity [8] or with CA as part of a multi-enzyme microbead where CO 2 capturing and subsequent processing take place in the same spot [9].
CAs are distributed over all forms of living organisms [10] and are essential for the transfer of CO 2 and bicarbonate (HCO 3 − ) by catalyzing the reversible hydration of CO 2 : In detail, the water molecule attached to the zinc atom of the active center of the CA is deprotonated and initiates a nucleophilic attack. The target is the carbon atom of CO 2 and the product is HCO 3 − which is liberated in exchange to another water molecule in the last step [11]:

PRACTICAL APPLICATION
The increasing amount of CO 2 in the atmosphere is a major source for the climate change and global warming. At the same time, it is a huge reservoir of one-carbon molecules. This potential has been addressed in the last years by setting up a roadmap to a biotechnological C1 economy. While the upcycling to valuable platform chemicals and fuels is making progress, an efficient CO 2 sequestration remains a critical issue. In this study we investigated a combined approach. Carbonic anhydrase (CA) is used to sequestrate CO 2 and catalyze the hydration into reactive HCO 3 − . The CA cahB1 from the alkaliphilic Coleofasciculus chthonoplastes was tested for this application. Additionally, the partial pressure of CO 2 is increased to 20 bar to maximize the productivity. In this way, larger amounts of CO 2 can be utilized once processes are scaled-up.
An elegant way to sequestrate CO 2 is to precipitate calcium carbonate (CaCO 3 ) out of an aqueous solution containing Ca 2+ ions. The solubility of CaCO 3 in water is low and while reducing the amount of CO 2 , synthetic precipitated calcium carbonate (PCC) is a product used in various applications as filler material [12] pharmaceutical carrier [13], or nutritional supplement [14]. One reason for the broad application is its polymorphism. Within the different pure crystalline structures as calcite, vaterite and aragonite, many shapes as plates, rhombohedra, needles and spherulites are possible. In nature a great variety of organisms use biomineralization to build parts like shells and other structures [15,16] Especially in corals and sponges calcium carbonate is a major material [17,18]. For invertebrates and vertebrates CA plays important role in biomineralization. Besides being part of respiration and acid-base balance processes [19], CA increases the calcification rate and manipulates the morphology. The whole mechanism is not completely understood but the potential motivates recent research [20,21]. Organisms seem to control precipitation to calcite [22], vaterite [23] or aragonite [24] via CA.
For the precipitation, an alkaline pH is mandatory in consideration of the carbonic acid equilibrium system [25], because HCO 3 − provided by CA can dissociate into carbonate ions (CO 3 2− ): In the next step, if calcium ions (Ca 2+ ) are present, they form CaCO 3 together with CO 3 2− .
Enzyme-enhanced CO 2 sequestration has been investigated in multiple manners: either by purifying and evaluating CAs from different origins [26,27] or by immobilizing CA to make it more resistant to harsh conditions [11,28]. Few studies dealt with the production of PCC under pressurized CO 2 . Montes-Hernandez et al. [29] suggested an increased precipitation rate due to the increased solubility of CO 2 while stating a limited impact of molecular CO 2 on the precipitation itself.
To the best of our knowledge, these two approaches, the enzymatic and the pressure induced accelerations of CO 2 sequestration process, have not been combined yet. Therefore, this study aims to evaluate the possible capture rate of CO 2 by comparing the performance of two different CAs under different pressure levels. Because of the characteristics mentioned before, PCC is an interesting product itself. Consequently, the produced CaCO 3 morphologies are characterized using SEM and XRD and the phase interchange during the precipitation process is investigated.

Expression and purification of CA cahB1 from Coleofasciculus chthonoplastes
The gene cahB1 was obtained from the group of Dr. Kupriyanova from the Russian Academy of Sciences (Moscow, Russia) who discovered the CA in the alkaliphilic cyanobacterium C. chthonoplastes (ex-Microcoleus chthonoplastes) and fused it to thioredoxin [30] with a polyhistidin-tag at the N-and C-termini of the protein [31]. The amplified construct was cloned into the recombinant plasmid pET-32b(+) (Novogen) and transformed into E. coli strain BL21(DE3). For CA production, the E. coli cells were cultivated at 37 • C in LB-medium until OD 600 = 0.7 was reached and induced afterwards by adding 1 mM isopropyl thio-b-d-galactoside (IPTG). The overproduction took place for 4 h at 25 • C with a final OD of 3. The cells were separated from the fermentation culture broth by centrifugation and homogenized with a Spectronic SLM Aminco French pressure cell. The expressed proteins were purified by affinity chromatography using a Knauer FPLC equipped with a HisTrap HP column containing nickel ions resins purchased from Sigma-Aldrich according to the manufacturer's protocol. Finally, the elution buffer was exchanged to purified water three times by recharging Amicon Ultra centrifugal filters with a molecular weight cut-off of 30 kDa.

Protein determination
The protein concentration was estimated performing the Bradford Coomassie brilliant blue assay [32] using bovine serum albumin as a standard and measuring absorbance at 595 nm.

Activity assay
The enzymatic activity assay of the alkaliphilic CA was performed according to Wilbur and Anderson [33]. In this assay CO 2 is used as a substrate. Specifically, a CO 2 saturated ice-cold solution is prepared by bubbling pure CO 2 into purified water. During the assay, the reaction vessel is tempered at 0 • C. To start the reaction, 20 mL of ice-cold CO 2 -saturated water is given into 30 mL of a 20 mM Tris-HCl mixture containing 15 μL purified water as a blank or 15 μL of enzyme solution. The Tris-HCl buffer was adjusted to pH 8.3 at room temperature and then cooled down to 0 • C. Wilbur-Anderson units (WAU) are defined as the ratio between the time required to drop the pH from 8.3 to 6.3 for the enzymatic test subtracted from the blanks time T 0 and divided again by the tests time: WAU is the standard unit for CA measured at 0-4 • C. All experiments were performed in triplicates.

Sequestration of CO 2 into CaCO 3 at ambient pressure
The ammonium carbonate diffusion method was adopted from Müller et al. [27] The carbonation of Ca 2+ was performed at ambient pressure in a desiccator. In this method, CO 2 is generated from ammonium bicarbonate (NH 4 HCO 3 ) solution. The upper compartment of the desiccator contained beakers with 10 mL of 50 or 100 mM CaCl 2 solution which was buffered to pH 9 with 25 mM Tris-HCl. The experiments were executed either without enzymes or with 2 WAU/mL of BCA or cahB1. The desiccator was placed on a benchtop shaker with a frequency set to 50 rpm. Triplicates were run at room temperature for different time spans up to 5 h. The reactions were stopped

Sequestration of CO 2 into CaCO 3 at increased pressure
Pressurized carbonation was performed in a high-pressure reactor system of the Parr Instrument Company with an internal volume of 300 mL. A scheme of the set-up is depicted in Figure 1. In parallel to the experiments conducted in the desiccator, 96 mL solutions of the same samples in terms of concentrations were prepared. The reactor was stirred at 50 rpm at room temperature. To start the precipitation, CO 2 was injected into the system. CO 2 was purchased from Westfalen Gas, Germany, with a purity of 99.99%.
The pressure was set to be constant at 5, 10, or 20 bar, meaning that CO 2 adsorbed by the solution would be recharged to keep the pressure constant. The pressure is achieved by a gas cylinder with at least 50 bar CO 2 inside and the system is flushed to remove all air to ensure a pure CO 2 phase. Subsequently, the total amount of CO 2 injected per run is the sum of CO 2 in the gas phase, in the solution and precipitated as CaCO 3 at the end of the run. The amount of CO 2 in the gas phase can be calculated by the Peng-Robinson equation of state [34], the solubility of CO 2 is calculated later in Section 3.3 and the amount of CO 2 in CaCO 3 equals the initial concentration of Ca 2+ . For the runs with 50 mM CaCl 2 this leads to 0.215 mol CO 2 injected at 5 bar, 0.42 mol at 10 bar and 0.8 mol at 20 bar. Each run was repeated three times.
In order to determine the amount of precipitated CaCO 3 and its characterization, samples of 1 mL were taken from the on-going process through a valve. Depending on the pressure and the initial Ca 2+ concentration the carbonations were run for up to 15 min until the reaction finished. The samples were treated as described before.

Determination of free Ca 2+
To follow the precipitation of CaCO 3 quantitatively, the concentration of the free Ca 2+ in the supernatant is determined by ethylenediaminetetraacetic acid (EDTA) titration. In this complexometric titration, the endpoint is detected by a color change of the indicator Eriochrome Black T due to a lack of Ca 2+ caused by the formation of Ca 2+ -EDTA complexes [35]. Subsequently, the concentration of Ca 2+ can be calculated using the volume of the supernatant V Ca 2+ , the concentration and volume of the EDTA solution c EDTA and V EDTA , respectively: The conversion of free Ca 2+ to CaCO 3 can be calculated using the concentration of free Ca 2+ in the beginning c 0 and the determined concentration at a certain time c t .

Characterization of formed CaCO 3
The morphology of the dried CaCO 3 was investigated using a scanning electron microscope (SEM) DSM 962 from Zeiss, Germany, operating at an accelerating voltage of 10 kV. The samples were suspended in ultrapure water, placed on a holder, dried again and sputtered with a layer of gold and palladium. The qualitative phase analysis was performed by using X-ray powder diffraction (XRD) measurements. A D8  [31]. Lane B resolves cahB1 purified by affinity chromatography. Both lanes were loaded with a total protein mass of 10 mg Endeavor diffractometer from Bruker, Massachusetts, USA, was utilized to determine the crystalline structure using Cu Kα radiation (λ = 15,406 Å) and a 2 Theta angle ranging from 4 • to 65 • . The quantitative phase analysis was done by the Rietveld method [36] using Topas from Bruker AXS.

Activity of cahB1
The successful production and purification of cahB1 was confirmed by electrophoretic analysis using SDS-PAGE under denaturing conditions ( Figure 2). The specific activity of the purified enzyme was determined to be 72.

CaCO 3 precipitation accelerated by carbonic anhydrase and pressure
In general, the precipitation rate of PCC depends on the concentrations of the starting compounds Ca 2+ and CO 2 [37,38]. To track the progress of the precipitation in different experiments, the removal of Ca 2+ is plotted for the experiments with 50 mM (A-D) and 100 mM (E-H) CaCl 2 solution in Figure 3. Please note the change in the scale of the x-and y-axis, for example it took roughly 2 min until no Ca 2+ was left starting at 50 mM at 20 bar (3D) compared to 4 min at 100 mM and 20 bar (3H).
The time required (t 80 ) for a removal of 80% of Ca 2+ or 80% precipitation of CaCO 3 was calculated using the exponential equation suggested by Stocks-Fischer et al. [39] to describe microbial CaCO 3 precipitation: In the equation, Δc i is the difference between the starting concentration c 0 and the final concentration of c Ca 2+ , k the reaction rate, t the time and z the time point of the maximum of (dc/dt). The parameters k and z can be derived by fitting Equation 10 to the experimental results of Figure 3 and were determined using OriginLab (Version 2020) in this work.
c Ca 2+ equals 0.2 since 80% of the Ca 2+ are removed, c 0 is 1 and Δc = c 0 , because the reaction runs until all Ca 2+ are removed, t 80 can be calculated by rearranging Equation 10: Table 1 shows the derived values of the parameters obtained from fitting the data of Figure 3 as described before. Generally, the greatest jump takes place between 1 and 5 bar where k increases by a factor of 10 to 30, while t 80 is decreased by a factor of 20 to 40. Afterwards, the increased pressure seems to lead to a proportional or reverse proportional change of the parameters, respectively. Interestingly, at a concentration of 100 mM CaCl 2 no catalytic effect is visible since the parameters are similar to the ones calculated without any enzyme. A possible explanation can be an inhibition caused by the Cl − [40]. In the next section, t 80 is used to calculate the sequestration rate of CO 2 .

Sequestration of CO 2 into CaCO 3
The improvement of the CO 2 sequestration due to an enzymatic catalysis and an increased pressure was examined. Table 2 shows the production rate of CaCO 3 and subsequently the rate of sequestered CO 2 . The results of this study refer to the experiments with an initial concentration of 50 mM CaCl 2 . For the calculation of the production rate, the starting concentration of CaCl 2 plays a minor role. However, a broad range of anions and other small molecules seem to inhibit cahB1 including chloride [41]. In fact, the different studies must be compared carefully. In this study and in the ones of Molva et al. [11] and Müller et al. [27], the solution is aerated across the surface while in the study of Montes-Hernandez et al. [29] the gas is bub-bled in and in the study of Chafik et al. [26] a solution is saturated with CO 2 and afterwards mixed with a solution containing Ca 2+ .
As expected, the production rate of CaCO 3 increases with the pressure according to Fick's law since the adsorption and hydration of CO 2 were identified as the TA B L E 1 Parameters obtained from fitting the data of Figure 3 to the Stocks-Fischer equation The diffusion flux can be calculated using C * L, CO2 the saturation concentration of CO 2 in the liquid phase corresponding to its partial pressure (p CO2 ) in the gas phase, C L, CO2 the real CO 2 concentration in the liquid phase, D L,CO2 the diffusion coefficient, the thickness of a "film" at the gas-liquid interface, and the area of gas-liquid interface. Assuming a pure CO 2 gas phase, C * L, CO2 can be calculated to be 33.
K H is the Henry constant taken from the PHREEQC data, CO 2 is the activity coefficient in water and φ CO 2 the fugacity coefficient. Note that CaCl 2 decreases the solubility of CO 2 in water which is reflected in CO 2 . For 100 mM CaCl 2 , CO 2 was found to be 1.07 [43]. Due to the model [44] applicable at concentrations below 3 M CaCl 2 , CO 2 can be calculated to be 1.05 using linear regression at 50 mM CaCl 2 . More recently, the solubility was modelled for higher temperature, pressure and CaCl 2 concentration [45] comparable to this study. The results show the expected impact at extreme conditions. The solubility increases approximately proportional within the partial pressure range of gaseous CO 2 used in this study. At higher pressure, the fugacity coefficient starts to have a substantial influence on the solubility and slows down its growth. In PHREEQC, the Peng-Robinson equation of state [34] is used to calculate the fugacity coefficient.
The production rate was calculated using the time t 80 derived before, 80% of the initial concentration of Ca 2+ TA B L E 2 Production rates from the carbonations of the 50 mM CaCl 2 solutions of this study compared to recent literature  2 in water leads to the formation of two hydroxide ions resulting in an alkaline environment. The pH value above 12.5 in saturated solutions decreases the activity of all common enzymes or even causes denaturation. Compared to the pH optimum around 8 to 9 of BCA [45], cahB1 seems to be adapted to the alkaline environment with a pH optimum around 9 to 10 [31] but did not show significant activity in preliminary experiments with Ca(OH) 2 .

Study p [bar] T [ • C] CA
On the other side, the pH of a CaCl 2 solution can be easily adjusted using a buffer. Interestingly, the supercritical CO 2 (90 bar, 90 • C) led to a lower production rate of 213.17 kg CaCO3 /m 3 h than the gaseous CO 2 (55 bar, 30 • C) with 789.6 kg CaCO3 /m 3 h. According to the authors, the lower gas solubility caused by the higher temperature is the reason for the reduced production rate of the supercritical CO 2 approach.
The approach Molva et al. is closer to the one executed in the desiccator. The Ca(OH) 2 solution is in contact with a CO 2 atmosphere through a defined gas-liquid interface. As a central aspect, free BCA was compared with immobilized BCA. The production rates of the free and immobilized BCA were 17.66 and 24.14 kg CaCO3 /m 3 h, respectively. In this study, production rates of 150.41 (without CA), 189.18 (BCA) and 243.68 kg CaCO3 /m 3 h (cahB1) were obtained at 1 bar. Remarkably, the fluxes through the interface in the study of Molva et al. and   The sample in A is taken 2.5 min after the beginning of the reaction and shows agglomerates of small particles. B (5 min. after beginning of the reaction) shows agglomerates of spheres mainly, with increased in size compared to A and can be identified as vaterite. After 10 min, a mix of spheric and rhombohedral crystals with a diameter around 5 μ can be found in C. In the middle panel, the results with 2 WAU/mL BCA are shown. The solid phase in D appears more mature compared to A since the particle size is increased and spherical vaterite is present. Smaller particles appear to have already a typical rhombohedral calcite shape. In E and F agglomerates of calcite crystals are dominant which increase in size during the time of the experiment. The SEM images on the bottom panel are taken from samples of the experiment with 2 WAU/mL cahB1. G is comparable to the small agglomerates of image A while H shows calcite particles with some remaining spherical vaterite. After 6 min, I shows again the final calcite particles distinctively stepped at the edges sponge CA, respectively. In this case, the lack of an agitation is most likely the reason for the lower production rates but the influence of the enzyme concentration is demonstrated. The increase of the CO 2 sequestration rate due to the presence of CA is 66% for 3 WAU/mL in the study of Müller et al. As mentioned before, in this study 19% for BCA and 60% for cahB1 were achieved. Note that only 2 WAU/mL were used, the results are again in a similar range. Chafik et al. investigated a camel liver CA, resulting in 6.76 kg CaCO3 /m 3 h. The experimental setup deviates but a high sequestration capacity was achieved compared to other works studying mammalian CA. Therefore, the results are included to give an overview.

Characterization of formed CaCO 3
To investigate the formation of CaCO 3 , the solid phases were characterized by using SEM and XRD. Due to similarity of CaCO 3 particles from 50 mM to CaCO 3 particles from 100 mM CaCl 2 solution regardless of a pressure of 5, 10 or 20 bar, not all samples are analyzed and discussed. The SEM images of the trials at 10 bar and 100 mM CaCl 2 are shown in Figure 4. Studies dealing with CA involved in the CO 2 sequestration and CaCO 3 precipitation focus on the morphology of the final product after the whole precipitation process is complete. In this work, the interchange of crystalline characteristics by taking and analyzing samples In the absence of CA, undefined slightly agglomerated particles with a size below 1 μm are present after 2.5 min of reaction time ( Figure 4A). Figure 4B shows grown particles after 5 min which start to look like spherical vaterite covered with a "dusty" layer. After 10 min ( Figure 4C) spherical vaterite particles are grown to up to 5 μm. In the classical approach nucleation takes place in a supersaturated ionic solution where meta-stable clusters are formed and decomposed, making the creation of a crystalline precursor which overcomes the critical cluster size a stochastic event. Modern theories favor a pathway of stable prenucleation clusters of ions, which can appear in undersaturated solutions as well, leading to an amorphous phase [46]. In the further course, the amorphous phase can translate to one of the water-free phases aragonite [47], vaterite or calcite [48]. Therefore, the observable structures in the beginning may be the result of an on-going interchange of amorphous calcium carbonate (ACC) to vaterite. A study in a similar system promotes a phase change from ACC to vaterite starting within the first minutes depending on the temperature [48]. The increased reaction speed due to the pressure supports the formation of ACC but at the same time a pure ACC phase is unlikely to observe. At the end ( Figure 4C), the interchange to calcite rhombohedra is taking place [49]. The finale state can be proven by the XRD pattern of the sample ( Figure 5A) [50] In the presence of 2 WAU/mL BCA ( Figure 4D), spherical vaterite with a diameter of 5 μm appears earlier at 2 min. According to the results described before, the precipitation reaction seems to be accelerated as well as the crystalline phase transformation. Again, the smaller particle agglomerated might be a late stage of ACC, in this case the cubic shape could refer to calcite. After 4 min ( Figure 4E) the conversion to calcite is already finished and until 8 min ( Figure 4F) the agglomerates grow to a size of around 5 μm.
In the corresponding XRD pattern ( Figure 5B) mere calcite phase diffraction peaks appear as expected.
The third panel shows PCC performed in the presence of 2 WAU/mL cahB1. In the beginning ( Figure 4G), a solid phase like the one in the beginning of the one in absence of CA is present. As mentioned before, it might be an amorphous precursor of calcite and/or vaterite. In Figure 4H rhombohedral calcite is the dominant phase with only a few spheric vaterite particles left. Again, this observation supports the assumption of a fast, morphologic interchange due to the present of carbonic anhydrase. After 6 min, a pure calcite phase is visible ( Figure 4I) proofed by the pattern of the XRD measurement ( Figure 5C). In this SEM image the edges of the cubic-like calcite particles appear more stepped compared to the shapes discussed before. Functional groups of proteins are expected to inhibit the growth of calcite at increased concentrations [51] and similar effects of shrunk edges can be found in many studies focusing on the biomineralization [52][53][54]. However, since the enzyme concentration is low in this study, the effect plays a minor role.
Former studies suggested a change from vaterite to calcite by a dissolution-reprecipitation aging mechanism [30,33]. The first generation of particles is (partly) dissolved and reprecipitated again to new more stable morphology. In the present composition, the higher solubility of vaterite compared to calcite [48] is the driving force in the process at ambient conditions. Therefore, in the final phase of the precipitation the solution is supersaturated regarding calcite but not vaterite. Subsequently, the phase shifts from vaterite to calcite with time. The mechanism supported by this study is highlighted in Figure 6. The high pressure and the catalysis of CO 2 to HCO 3 − by CA leads to a highly supersaturated solution and a fast precipitation rate. As mentioned before, ACC is expected to be part of the process. The packing density of Ca 2+ in AAC is similar to the one in crystalline forms like vaterite and calcite, making an interchange to this phases possible [55]. In the three SEM images the dissolution-reprecipitation aging is illustrated. The surface of the vaterite is dissolved and reprecipitates as calcite. This process continues in aqueous solution, until a pure stable calcite phase remains.
In contrast to the SEM images from samples taken from the Parr-reactor, the particles obtained from the precipitation in the desiccator did not show a systematic pattern.
The solid phases in Figure 7 look like typical calcite rhombohedra which are magnitudes larger than the particles found in the Parr-reactor. A cause might be the longer reaction time, resulting in the conversion into calcite before the first sample was taken. Furthermore, the aging and growing of the particles is more present because the slight supersaturation leads to stable crystals. On the other side, the high supersaturation of the pressure experiments leads to the rapid nucleation of multiple precursors. It is interesting to note that many agglomerates in Figure 7C show a smooth flat surface on one side, indicating that they were located at the glass wall of the beaker or at the gas-liquid interface. This was not visible in the Parr-reactor samples, consequently the fluid mechanics and therefore the mass transport should be compared carefully between both approaches.
In addition, some flower-like crystallographic structures ( Figure 7B) appeared in the desiccator experiments. This unusual phenomenon does not seem to be an artefact since fragments of similar flowers can be found in 6A highlighted by the white circles. In fact, these agglomerates were described before [56]. They were found when precipitating CaCO 3 on the inner surface of an eggshell membrane. The composition of a vaterite sphere in the middle surrounded by calcite petals was suggested by Takiguchi et al. and is supported by the SEM images obtained in this study. The structure might be the result of a covering process leading to capsulated vaterite particles.

CONCLUDING REMARKS
For the first time, the approach of using CA for biological CO 2 sequestration was performed in a high pressure reactor with up to 20 bar. Additionally, a novel CA (CA cahB1) from the alkaliphilic cyanobacterium C. chthonoplastes was shown to be a promising enzyme showing a higher CO 2 sequestration rate than the mostly used BCA. An explanation is the difference of the pH tolerance of both enzymes. The adaptation to the alkaline environment of cahB1 is advantageous even if the physiological reason is not discovered yet. Since the precipitation of CaCO 3 highly depends on the presence of CO 3 2− a more acidic pH would not increase the total reaction rate of the CO 2 sequestration. The highest CO 2 sequestration rate of 2682.88 kg CaCO3 /m 3 h for cahB1 was achieved at 20 bar. The rate is higher than the ones reported so far using CA or pressurized CO 2 alone. Interestingly, the enhancement of CO 2 sequestration of the CA decreased with increasing pressure, in case of cahB1 from 60.1% at ambient pressure to 20.6% at 20 bar. Subsequently, further increased pressure can lead to even higher sequestration rates but demands for higher enzyme concentrations. The next step is to develop suitable devices which can applicate the CO 2 sequestration technology and link it to additional reactions for upcycling. The increased process costs associated with "single use" of purified enzymes may be addressed using approaches like enzymatic liquid membranes and other immobilization techniques. It is understood that cost efficiency is an economic challenge of enzyme-aided CO 2 sequestration.
Additionally, the precipitation of CaCO 3 under pressure and influence of CA was investigated and an interchange of the crystalline phase from vaterite to calcite was observed in detail. As a precursor, amorphous calcium carbonate is expected to play a major role in highly supersaturated solutions but could not be identified without doubt. Future work should focus on the very beginning of the precipitation.

A C K N O W L E D G M E N T S
The authors acknowledge support by the project consortium protPSI (protein pressure specific activity impact) funded by the German Ministry of Education and Research (031B0405A) and its partners, especially Prof. Andreas Liese and coworkers from the Institute of Technical Biocatalysis at the Hamburg University of Technology.
Open access funding enabled and organized by Projekt DEAL.

C O N F L I C T O F I N T E R E S T
The authors declare no conflict of interest.

D ATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.